Anuradhapura: The British-Sri Lankan Excavations at Anuradhapura Salgaha Watta 2. Volume I: The Site 9781841710365, 9781407351575

Citation preview

BAR  S824  1999  

Society for South Asian Studies Monograph No 3

CONINGHAM  

Anuradhapura The British-Sri Lankan Excavations at Anuradhapura Salgaha Watta 2

ANURADHAPURA

Volume I: The Site

VOLUME I: THE SITE

Robin Coningham

BAR International Series 824 9 781841 710365

B A R

1999

Society for South Asian Studies Monograph No 3

Anuradhapura The British-Sri Lankan Excavations at Anuradhapura Salgaha W atta 2 Volume I: The Site

Robin Coningham with contributions from

F .R. Allchin, Cathy Batt, Paul Cheetham and Randolph Haggerty

BAR International Series 824 1999

Published in 2016 by BAR Publishing, Oxford BAR International Series 824 Society for South Asian Studies Monograph 3 Anuradhapura

© R Coningham and the Publisher 1999 The author's moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher.

ISBN 9781841710365 paperback ISBN 9781407351575 e-format DOI https://doi.org/10.30861/9781841710365 A catalogue record for this book is available from the British Library BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 197 4 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by Archaeopress in conjunction with British Archaeological Reports (Oxford) Ltd/ Hadrian Books Ltd, the Series principal publisher, in 1999. This present volume is published by BAR Publishing, 2016.

BAR

PUBLISHING BAR titles are available from:

EMAIL

PHONE FAX

BAR Publishing 122 Banbury Rd, Oxford, OX2 7BP, UK [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

!l§·i: i-'□··o., ··:_·; ....~·.:·· ....··. ft:

•

. .•

for Paula and Urban, in lieu of lost days

V

CONTENTS

F.R. Allchin

Preface Acknowledgements List of mustrations

ix xiii xv

1

Introduction

Robin Coningham

1

2

The PhysicalEnvironment 2.1 Introduction 2.2 Geology 2.3 Climate 2.4 Drainage and relief 2.5 Flora 2.6 Fauna 2.7 Soils

Randolph Haggerty and Robin Coningham

7 7 7 9 9 10

The City of Anuradhapura 3.1 Introduction 3.2 History of archaeological research 3. 3 The Citadel zone 3.4 The monastic zone 3.5 The tank and village zone 3.6 The forest and hermitage zone 3.7 Conclusion

Robin Coningham

3

4

s

11 11

15 15 15 17 21 23 25 26

The Fortifications Robin Coningham and Paul Cheetham 4.1 Introduction 4.2 Previous investigationsof the Citadel's defences 4.3 Methods and techniques 4.3.1 Surface survey 4.3.2 Proton magnetometer survey 4.3.3 Earth resistivity survey 4. 3 .4 Soil auger coring survey 4.4 The northern fortifications 4.5 The eastern fortifications 4.6 The southern fortifications 4. 7 The western fortifications 4.8 Dating the fortifications 4.9 Conclusion

47 47 47 48

The Excavationsat ASW2 5 .1 Introduction 5.2 Structural period K 5.3 Structural period J 5. 4 Structural period I 5.5 Structural period H 5. 6 Structural period G 5. 7 Structural period F 5. 8 Structural periods D and E 5. 9 Structural period C 5 .10 Structural period B 5 .11 Structural period A 5.12 Conclusion

71 71 71 72 74 77 77 79 80 80 81 82 82

Robin Coningham

vii

48 48

48 49 49 50 51 51 52 54

Anuradhapura:The Site 6

Dating the Sequence

Robin Coningham and Cathy Batt

6.1 Introduction 6.2 Structural period K 6.3 Structural period J 6 .4 Structural period I 6.5 Structural period H 6.6 Structural period G 6. 7 Structural period F 6.8 Structural periods E, D, C, and B 6.9 Structural period A 6.10 Conclusion

7

125 125 126 126

127 128 128

129 129 130 130

Conclusions

Robin Coningham

135

References

149

AppendixA: Sri Lanka's Rulers: A ChronologicalList

155

AppendixB: ASW2 Context Descriptions

159

AppendixC: RadiocarbonAges of MeasuredSamples from ASW2

209

viii

PREFACE F.R. Allchin

Moreover, the proximity of their ashrams to major settlements seemed to be a requisite for the sages to enjoy the material support of the local population. Surely, I argued, this indicated that the sages and disputants in the BrihadaranyakaUpanisadand the ChandogyaUpanisad were not the forerunners of the early cities, but rather among their products. Regarding the rise of Buddhism the case was even clearer, since it was generally agreed that this took place in the context of the emergence of a powerful merchant class in the cities. Moreover, the Suttas themselves gave ample support for the close links between early Buddhism and the cities. In those days I was unaware of Max Weber's Hinduismus und Buddhismus(1920, English translation 1958), where this view had been expressed with great clarity some forty years earlier. My attention was drawn to early city formation in Sri Lanka during several visits to that country in the late 1980s. On one of these I read a paper, later published in 1989 in South Asian Studies (Allchin 1989), arguing that city formation appeared to have taken place more or less contemporaneously throughout South Asia. If this were the case, it should be treated as a South Asia-wide process and not thought of in isolation in any one part of the subcontinent. These occasions also offered an opportunity to visit the Citadel mound at Anuradhapura and witness at first hand the excavation of the carefully planned and sited sondages which Dr Siran Deraniyagala of the Department of Archaeology was at that time directing. I was struck by his bold and original plan, which coincided with my belief in the importance of purposefully designed research projects. I was particularly impressed by his systematic use of radiocarbon dates to provide a proper archaeological chronology (something which was surprisingly rare in most South Asian excavations before this). If this were combined with his carefully excavated stratigraphic sequence, a solid basis for chronology must emerge. This matter became even more important when early Brahmi inscriptions, scratched on pottery, were reported from the sondages. At first I was frankly sceptical and expressed the view that the inscriptions must be much younger than the radiocarbon dates suggested. All these things highlighted the need for such problem-oriented projects concerning the Early Historic period throughout South Asia. Until this wish becomes a reality our knowledge of the early cities and states of India, Pakistan, Afghanistan and Nepal must remain fragmentary and lack the balanced perspective that archaeology can provide. While appreciating the originality of the concept of Dr Deraniyagala's project, I also became aware of its

THE origins of the cities of Greece and Rome have long been the subject of scholarly attention, together with the social evolution that accompanied the formation of their states. In the mid-nineteenth century ancient historians, and particularly Fustel de Coulanges, achieved a striking depth of understanding of these themes, almost wholly on the basis of textual data (Coulanges 1864). This was long before archaeologists developed an active interest in such things. Historians of South Asia were slow to follow in their footsteps. Only in the past half century have such scholars as D .D. Kosambi, Debriprasad Chattopadhyaya, Romila Thapar and Ram Sharan Sharma begun to focus attention on what was involved in the rise of cities in South Asia and in its social and economic implications. With the notable exception of Marshall's excavations at Bhita (1912) and Taxila (1951), South Asian archaeologists for the most part have neglected to use their discipline to contribute to the study of Early Historic cities. A. Ghosh in his small book The City in Early HistoricalIndia (1973) and D.K. Chakrabarti in his PhD thesis on Early Urban Centresin India (1972) were each in their own way pioneers. Nevertheless, up to the present time, archaeological fieldwork at Early Historic cities in India has been largely limited to cutting sections through city ramparts with a view to establishing their chronology, and the unique opportunities offered for example by the Nagarjunasagar dam project - have to a large extent been wasted. My introduction to the early cities of the Ganges valley goes back to the early 1950s. At that time my wife and I were inspired by a number of site visits. These included Hastinapur in the company of B.B. Lal and B.K. Thapar; Kausambi under the guidance of Professor G.R. Sharma, then excavating there; and Kumrahar, where we were happy to meet Vijayakanta Mishra, a brilliant young archaeologist whose early death robbed Indian archaeology of a most promising career. These visits gave me the beginnings of a new perspective which was far removed from the text-centred history I had hitherto learned. I remember giving a lecture in London around that time in which I argued that the society which gave rise both to the early Upanisads and to Buddhism should be seen in the context of the life and culture of the new cities. With regard to the former philosophical texts, forest hermitages may have been their setting, but they must also have been near enough to cities to permit the meetings and discussions with local kings (such as Ajatasatru of Kasi and Janaka of Videha) and the bestowal of royal gifts of great herds of cattle. ix

Anuradhapura:The Site excavation results. No one will dispute that preliminary reports should be published as early as possible. But the preparation of a final report, particularly when it has involved several seasons' work, must demand sufficient time for adequate research and study, for technical and scientific analyses to be made, and for the report itself to be written and edited. A few years' delay is a small price to pay for getting a comprehensive and maturely produced report. Surely this is to be preferred to a hastily written report, lacking in both depth of research and scholarship? Clearly there is a happy mean between too much haste, on the one hand, and too much perfectionism or delay on the other. One regrets that some grant-giving bodies seem to regard it as their duty to harass recipients of grants, particularly when the latter happen to be relatively junior, even threatening them, to produce final reports with unreasonable speed, as though speed were the overriding consideration. This it should never be. The Sri Lankan-British excavations at Anuradhapura, site ASW2, constitute a further contribution to the growing body of data regarding the early history of the site. The present publication is in two parts, that is, two volumes, the first dealing with the site, the excavations and stratification, and the second with the various categories of objects and materials discovered. The chronology of the early occupation emerges with considerable clarity and, in spite of minor differences between our dating and that obtained from Deraniyagala's sondages, the impressive number of radiocarbon samples deriving from both must make Anuradhapura archaeologically one of the most firmly dated Early Historic cities of the subcontinent. The radiocarbon chronology can be applied with confidence to the remarkable series of structural periods discovered in the site and discussed in chapters 5 and 6 of the present volume. The earliest structures, in periods K and J, were circular huts with timber posts and wattle and daub walls, and buildings of this type continued to be made through a series of eight or more reconstructions. Structural period I witnessed a major change with the introduction of rectangular buildings of timber and wattle and daub, apparently carefully oriented. This was evidently the time when Anuradhapura was refounded as a city planned on traditional South Asian lines. Once again in this period there were around eight further reconstructions. A substantial change in building construction and materials took place in structural period G, when solid buildings of burnt brick and occasional limestone blocks, and terracotta roof tiles, appeared for the first time. This period too witnessed a further five reconstructions. From structural period F forwards one more major change in building materials was introduced, with structural stonework, including columns and beams, used to a quite unprecedented extent. This period heralded the beginning of the largely stone architecture which was to remain a feature of the 'Anuradhapura period', throughout its life as a capital city. Thus in the course of the excavations at ASW2 well over a millennium of structural history has been identified and dated.

limitations. The accumulated deposits of the Citadel mound at Anuradhapura are around ten metres in depth, and the excavation of sondages of only three metres square to such a depth was something of a tour de force. To begin with, working in so small an area makes the task of accurately recording the stratigraphy very difficult to achieve, not to say hazardous. For example, how accurate a picture can be obtained of features such as pits, or portions of pits? Moreover, the excavation of single sondages of such dimensions more or less completely rules out the possibility of discovering identifiable rooms or other structural features, let alone complete houses. Again, the opportunities for recording the positional contexts of finds in relation to other objects or features must be reduced by the small area available, and many other aspects of cultural data are likely to be inaccessible. To me it appeared that a logical next step would be to undertake the excavation of a larger area, making possible the observation of areas of actual occupation large enough to reveal such evidence. Dr Deraniyagala told me that he would welcome collaboration from foreign excavators. I suggested that we would like to field a British team and he expressed a positive interest. I added that, on account of my age and imminent retirement, I should want to entrust the field direction to a younger person. Further discussions of this plan were held with Dr Roland Silva, then Director of Archaeology, Professor Senaka Bandaranayake and others. The outcome was that Dr Deraniyagala very kindly invited our team to work under the aegis of his major Anuradhapura project and generously offered us the ASW2 site, alongside his own ASW sondage. The outcome was the six seasons of excavation and postexcavation fieldwork carried out between 1989 and 1994. I would like, at this point, to express my gratitude and that of the rest of our team to Dr Deraniyagala and his colleagues in the Sri Lankan Archaeological Department for their continuing support and encouragement. Without this our project would have been difficult, if not impossible to achieve. I am happy therefore to be invited to write the Preface to the report of the excavations at ASW2. I was present at the first conception of the project, and so it is a matter of great satisfaction to see it reach its conclusion. The completion of the final report and its publication marks the fulfilment of the obligation we accepted from the start. To reach this point within a decade of the inception of the project is quite an achievement, bearing in mind the number of field seasons involved and the other preoccupations of several of the chief actors. No sooner was a field season completed than normal duties recommended, including not least - the earning of their livings. The successful completion of the several specialist reports from authors who were also heavily committed in other areas deserves recognition! Perhaps this is a good point at which to make a small digression and express the personal view that all too often these days over-much emphasis is placed upon the need for prompt, almost instantaneous, publication of

X

Preface throughout structural period J and onwards, only disappearing around the start of the Christian era. The first examples of crudely scratched Brahmi letters occur before the end of period J, and from period I onwards more regular use of script is found. The short inscriptions on pottery appear to have served as a means of denoting ownership, either of the vessel or of its contents. The graffiti appear first in period K, while the inscriptions occur first at the end of period J. Thereafter both systems run parallel to one another. We may infer that both were used for the same purpose, to identify ownership, either of an individual or of a family or clan. The script shows only slight developments during the following two or three centuries, but virtually none of the changes of script which occurred in Sri Lanka in the early centuries AD are present, suggesting that the custom of inscribing pottery with owners' marks or inscriptions ended before that time. By c. 200 BC some of our inscriptions bear the names and titles of several royal ladies and an official, all of whose names are already known from early Buddhist cave inscriptions. In sum, the excavations at Anuradhapura provide a wonderful database of evidence relating to the early history of Sri Lanka before, at and after the time when it became a capital city. From it we can study the stages of the emergence of a city and its subsequent growth. The record presented in this publication and its companion volume is, as far as can be, honest and objective. A point is now being reached at which one may confront the early textual tradition of Sri Lanka with the archaeological data. The study of the evidence for trade and foreign imports takes us far beyond the island; while the inscriptions and graffiti take us into the very centre of the state, maybe even into the residence of the ruling family. The excavations leave many questions unanswered and raise many new and difficult problems of interpretation. But they do, we believe, point to the way in which archaeology cancontribute to advancing knowledge of the processes of early city and state formation in South Asia.

Our work was also able to throw new light on the nature and history of the fortifications surrounding the city. We were fortunate to be able to study the section excavated in 1992 by the Sri Lankan-Japanese team, near the southern gate. We made a fresh study of the ditch and ramparts from the surface of the mound by remote sensing, using a proton magnetometer, a resistivity meter and a soil auger. By these means it was possible to obtain both a preliminary chronology for the city ramparts and a much wider perspective of the fortifications than could be had from the customary excavation of a single cross-section, used on many Early Historic ramparts in northern India. In the second volume of the report the finds are studied. These reveal much interesting and exciting material, of which one or two categories may suffice as examples. The pottery is dealt with in Chapter 5: Glazed Ceramics and Chapter 6: Unglazed Ceramics. The predominant pottery, from period K forwards, was a black and red burnished ware, made by a craft tradition and producing a range of forms strikingly similar to those reported in early Iron Age settlements throughout peninsular India. By structural period I the pottery range was augmented by two imported fabrics, a grey ware of a finely sorted clay and Rouletted ware of equally finely sorted clay. Along with these wares came other categories of finds providing evidence of importation and of growing foreign trade - pieces of glass, varieties of beads etc. From period I onwards small numbers of coins begin to appear in the excavations. These are studied in Chapter 2: Coins and, incidentally, provide an opportunity for a critical comparison of the chronologically determined stratification in relation to the various coin types. It may be remarked that in early South Asia coins are for the most part imprecisely dated by their find spots, while at the same time the chronology of many coin types has remained ill-defined. Hence there is a real need for this kind of analysis. The inscriptions and graffiti found in ASW2 are studied in Chapter 9: Epigraphy. Graffiti appear on pottery

xi

ACKNOWLEDGEl\'.IENTS

I am most grateful to the three Directors-General of Archaeology, Dr Roland Silva, the late Mr M. Sirisoma and Dr Siran Deraniyagala, who held office during our fieldwork at Anuradhapura between 1989 and 1994. Together, and singularly, they provided excellent support and collaboration for the team. I am also extremely grateful to Dr Roland Silva for his continued assistance as Director-General of the Cultural Triangle. Special thanks are, however, reserved for Dr Siran Deraniyagala, first as Director of the Anuradhapura Citadel Archaeological Project, and later as DirectorGeneral of Archaeology. The pioneer of scientific investigation at the Citadel of Anuradhapura, he acted as a mentor to the field team and was an exceptional source of knowledge about the archaeology of the Citadel and the island itself. Indeed, without Dr Deraniyagala's generous invitation to Raymond Allchin to mount an archaeological expedition to the Citadel of Anuradhapura, none of the fieldwork published here would have been undertaken. Dr Deraniyagala is internationally recognized as an expert on Sri Lankan prehistory, and the value of his contributions to the archaeology of Early Historic South Asia through his work at Anuradhapura cannot be over-estimated. The present volume builds on his pioneering work. Thanks must also go to the directors, officers and staff of the Cultural Triangle Jetavana and Abhayagiri projects in Anuradhapura, especially Dr Hema Ratnayake and Professor Hetterachchi. In addition, great assistance was provided by Dr Bridget Allchin, Dr Raymond Allchin, Dr Janet Ambers, Mr RobertJanaway, the late President J.R. Jayewardene, Mr Rukshan Jayewardene, Dr N. Kemp, Mr Nimal Perera, Dr Martha Prickett, Dr Sudashan Seneviratne, Dr Colin Shell, Professor van Andel and Mr Wijepala. In particular, I should like to warmly thank Rukshan and Viji Jayewardene for their hospitality, support and friendship over the years. A great debt of gratitude is owed to the members of the field teams of officers, students and staff from the Archaeological Survey Department, the Cultural Triangle and the universities of Bradford, Cambridge, Keleniya, Peradeniya, the Post-GraduateInstitute of Archaeological Research and Sri Jayewadenapura who worked at the site. Whilst there are too many to name individually, in particular I would like to thank the following: Dr Bridget

Allchin; Mr Kalum Nalinda Manamendra Arachchi; Claudia Beukroann, MA; Steve Cheshire, BSc; Masaki Choya, BA; Paula Coningham, MA; Gary Dooney, MA; Luxman Chandra, MA; Antonia Douthewaite, MA; Rukshan Jayewardene, MPhil; Dr Carl Knappett; Mr Alfred de Mel; Mr P.O. Mendis; Halawthage Jude Perera, BA; Mr P.R. Premachandre; Simon Weston, MA; and Sarah Wilde, BA. The Anuradhapura Citadel Archaeological Project lab teams provided excellentbackup and support, for which I am very grateful. I would also like to acknowledge the efforts of the lab teams and experts who have helped prepare the field data in the UK for publication: Dr Cathy Batt; Paul Cheetham, MPhil; Steve Cheshire, BSc; Dr Randolph Haggerty; Jon Sygrave, BSc; Ruth Young, MPhil; and Alistair Wilson, BSc. Finally, I should like to acknowledge the work of Jenny Marsh as copy editor and for seeing this volume through to press. The financial support of the following bodies is also gratefully acknowledged: the Archaeological Survey Department of Sri Lanka; the Ancient India and Iran Trust, Cambridge; the British Academy; the British Council, Colombo; Churchill College, Cambridge; King's College, Cambridge; the McDonald Institute for Archaeological Research, Cambridge University; the Overseas Development Administration; the Society of Antiquaries of London; and the Society for South Asian Studies (British Academy). I would also to thank the following institutions for temporary loan of their equipment for fieldwork: the Archaeological Survey Department of Sri Lanka; the Department of Archaeological Sciences, Bradford University; and the McDonald Institute for Archaeological Research and Department of Earth Sciences, Cambridge University. This research was carried out while the author held a college studentship at King's College, Cambridge, between 1989 and 1993; a research fellowship at the Ancient India and Iran Trust in Cambridge between 1993 and 1994; and a lectureship and senior lectureship in South Asian Archaeology at the Department of Archaeological Sciences, University of Bradford, from 1994.

LIST OF ILLUSTRATIONS

Alistair Wilsonkindly prepared Figures33, 37, 41, 45 and 53. The remainingplans wereprepared 1JyJohn Sigrave, the maps 1JySteve Cheshire,and the photographswere taken 1JyRobin Coningham. ColourPlates Ia

Conducting a geophysical survey of the Citadel's fortifications

xxi

lb

TrenchASW2

xxi

Ila

Buildings A and C with Daladage and Mahapali in background

xxii

IIb

The Mahathupa

xxii

IIIa

The Abhayagiri stupa

xxiii

Illb

Jaya Ganga channel

xx.ill

IVa

The Kiribat vihara stone bridge

xxiv

IVb

Exposed wall close to eastern gate

xxiv

Va

TrenchASW2

XXV

Vb

Structural phase J2

XXV

VIa

Pit 1371 (structural phase J3)

xx.vi

Vlb

Structural phase I1

xx.vi

Vila

Furnace or oven 1109 and 1111 (structural phase 12)

xx.vii

Vllb

Floor 613 and sf 10186 (structural phase G3)

xx.vii

VIIIa

Floor 408 (structural phase G4)

xx.viii

Vlllb

Floor 405 (structural phase G5)

xx.viii

IXa

Walls 339, 437, 442, 444, 453, 455 and 471 (structural phase G5)

xx.ix

IXb

Structural phase F

xx.ix

Xa

The megalithic cemetery at lbbankatuva

XXX

Xb

Puja around Buddhist stupa at Mihintale in 1990

XXX

Figures 1

Map showing location of Sri Lanka and Anuradhapura

5

2

Surveying structures in trench ASW2

6

3

Taking an auger core through the Citadel's archaeological deposits

6

4

Map showing Sri Lanka's topography and rainfall

13

5

Map showing Sri Lanka's raw materials

14

6

Plan of Anuradhapura

29

7

Schematic plan of Anuradhapura (after Seneviratna 1994)

30

8

Section of sondage AG-69 (after Deraniyagala 1972)

31

9

Plan of the Citadel showing the location of sondages

32

xv

Anuradhapura:The Site 10

The western defences of the Citadel showing ramparts on right and silted moat in foreground

33

11

Plan of final occupation phase street (after Ayrton 1924)

34

12

Plan of Paranavitana's excavations (after Paranavitana 1936)

35

13

Building A

36

14

The Gedige

36

15

Augering in front of 'Vijayabahu's palace'

37

16

Plan of 'Vijayabahu's palace'

38

17

The Bodhi tree shrine

39

18

The Thuparama stupa

39

19

The Mirisavati stupa

40

20

The Jetavana stupa

40

21

The Basavakkulam

41

22

The Tissavava

41

23

The Nuvaravava

42

24

The Toluvila stupa

42

25

The Western monasteries

43

26

The Vessagiri vihara

43

27

An early Brahmi inscription at Vessagiri

44

28

The Meghagiri vihara

44

29

The northern annicut on the Malvatu Oya in the process of being encased within a modem dam

45

30

The Kalavava

45

31

Plan of the Citadel (after Hocart 1924)

57

32

Auger core profile through the Citadel

58

33

Plan of the northern survey sector

59

34

The northern fortifications in situ

59

35

Resistance survey of the northern fortifications

60

36

Auger core profile through the northern fortifications

37

Plan of the eastern survey sector

61

38

The eastern fortifications

61

39

Resistance survey of the eastern fortifications

62

40

Auger core profile through the eastern fortifications

41

Plan of the southern survey sector

63

42

Resistance survey of the southern fortifications

64

43

Auger core profile through the southern fortifications

44

The southern fortifications

65

45

Plan of the western survey sector

65

46

The western fortifications

66

47

Section and auger profile through the southern rampart

67

xvi

facing page

facing page

facing page

60

62

64

List of Rlustrations 48

Primary rampart construction phase

68

49

Sixth rampart construction phase

68

50

View from the southern fortifications to the Mahathupa

69

51

Plan of ASW2' s immediate environment

85

52

Exposed pillar tops at ASW2

86

53

Southern section of ASW2

87

54

Plan of structural phase Kl

88

55

Plan of structural phase K2

88

56

Plan of structural phase K3

89

57

Structural phase K3 from the northeast

89

58

Well 1279 (structural phase K3)

90

59

Plan of stratigraphic phase XI

90

60

Plan of structural phase J1

91

61

Plan of structural phase J2

91

62

Plan of structural phase J3

92

63

Plan of pit 1371 (structural phase J3)

93

64

Plan of structural phase J4

94

65

Structural phase J4

94

66

Furnace or oven 1235 (structural phase J4)

95

67

Plan of structural phase JS

95

68

Plan of stratigraphic phase XXII

96

69

Plan of structural phase 11

96

70

Furnace or oven 1148 (structural phase 11)

97

71

Plan of structural phase 12

97

72

Plan of furnace or oven 1109 and 1111 (structural phase 12)

98

73

Plan of structural phase 13

99

74

Plan of structural phase 14

99

75

Structural phase 14

100

76

Plan of clay floors 972, 973, 974, 975 and 976 (structural phase 14)

100

77

Plan of tile collapse 894, 905, 912 and 914 (structural phase 14)

101

78

Tile collapse 894, 905, 912 and 914 (structural phase 14)

101

79

Plan of structural phase 15

102

80

Plan of structural phase 16

102

81

Plan of structural phase 16

103

82

View of vessel 879 in situ (structural phase 16)

103

83

Structural phase 17

104

84

Plan of structural phase 17

104

85

Plan of structural phase 18

105

86

Plan of trough 736 (structural phase Hl)

105

87

Plan of trough 733 (structural phase H2)

106 xvii

Anuradhapura:The Site 88

Plan of structural phase H 1

107

89

Plan of structural phase H2

107

90

Plan of structural phase O 1

108

91

Plan of structural phase 02

108

92

Plan of structural phase 03

109

93

Plan of structural phase 04

109

94

Plan of structural phase 05

110

95

Paving 450 (structural phase 05)

110

96

Foundation pit 669 and slot 637 (structural phase 05)

111

97

Plan of pillar foundations (structural phase F)

111

98

Pillar foundation 306 (structural phase F)

112

99

Pillar foundation 306 (structural phase F)

112

100

Elevation of pillar foundation 370 (structural phase F)

113

101

Plan of pillar foundation 358 (structural phase F)

114

102

Plan of pillar foundation 304 (structural phase F)

115

103

Plan of structural phases D and E

116

104

View of base of pit 535 (structural phases D and E)

116

105

Pit 274 (structural phases D and E)

117

106

Pit 274 (structural phases D and E)

117

107

Plan of structural phase C

118

108

Structural phase C

118

109

Plan of structural phase B 1

119

110

View of southeast corner of structural phase B 1

119

111

Plan of structural phase B2

120

112

Plan of structural phase B3

120

113

Plan of structural phase B4

121

114

Structural phase B4

121

115

Plan of structural phase B5

122

116

Plan of structural phase A

122

117

Plan of the surface at ASW2 prior to excavation

123

118

Sf 17420 (structural phase J4)

124

119

Phased and calibrated dates from ASW2

133

120

Map showing later prehistoric and Iron Age sites in Sri Lanka

143

121

Map showing Early Historic urban sites in South Asia (adapted from Allchin 1990)

144

122

Map showing an alternative location of Sri Lanka

145

Table Table 1 Chronological summary of the sequence at ASW2

xviii

xix

Table 1: Chronolo2ical summarv from the seauence at ASW2 (earliest ao1,earance) Structural phases A B5 B4 B3

B2 Bl C,D&E F

GS G4 G3 G2

~. >4

Gl H2 Hl 18 17 16 15 14 13 12 11

JS J4 J3 J2

Enameled sim (1918) and George VI coin (1943 AD)

{ c. 600 AD - 1100 AD Pandyan coin, Lustre ware, Lead Glazed ware, White, tin glazed ware, Buff ware, Sassanian-Islamic Blue glazed ware, Changsha painted stoneware, Yue green ware and Xing and Ding wares. Granite and Late Roman Imperial Third Brass Maneless lion coin, coooer alloy kohl stick, Eastern Mediterranean glass and blue glazed ceramics ICaitya& Fish coin Lakshmi plaque and copper alloy mirror Limestone & pillar foundations, Arikamedu type I 0, Nandipada & Swastika coin, Tree & Swastika coin, Tree & Caitya coin and iron saw Elephant & Swastika coin

{ c. 200 - 600 AD

{ c. 200 cal BC - 130 cal AD

Punch-marked coin and copper alloy vessel with laurel leaf design !Flap-shell turtle and palm fibre Mangrove wood species Lapis lazulin and Rouletted ware Bamboo Square structures, roof tiles & horse Marine shell (:amelian, marine turtle and early Brahmi scriptural ~affiti Rice Burial pit?, cuartz and Grey ware

{ c. 360 cal BC - 190 cal BC

{ c. 340 cal BC - 510 cal BC

Jl K3 { c. 460 cal BC - 840 cal BC

K2

Kl

Round structures? and iron objects

Colour Plates

Pl. Ia: Conductinga geophysicalsurvey of the Citadel'sfortifications

Pl. lb: TrenchASW2 xxi

Anuradhapura: The Site

Pl. Ila: Buildings A and C with Daladage and Mahapali in background

Pl . lib: The Mahathupa

xxii

Colour Plates

Pl. Illa : The Abhayagiri stupa

Pl. Illb: Jaya Ganga channel

xxiii

Anuradhapura:The Site

Pl. !Va: The Kiribat vihara stone bridge

Pl. !Vb: Exposed wall close to easterngate xxiv

Colour Plates

Pl. Va: Trench ASW2

Pl. Vb: Structural phase 12

XXV

Anuradhapura:The Site

f'

Pl. Via: Pit 1371 (structuralphase 13)

Pl. Vlb: Structuralphase II xxvi

Colour Plates

...

Pl. Vila: Furnace or oven 1109 and 1111 (structural phase 12)

Pl. Vllb : Floor 613 and sf 10186 (structural phase G3)

xxvii

Anuradhapura: The Site

Pl. Vllla: Floor 408 (structuralphase G4)

Pl. VII/b: Floor 405 (structuralphase G5)

xxviii

Colour Plates

Pl. IXa: Walls 339, 437, 442, 444, 453, 455 and 471 (structural phase G5)

Pl. IXb: Structural phase F

xxix

Anuradhapura: The Site

Pl. Xa: The megalithiccemeteryat Ibbankatuva

Pl. Xb: Puja aroundBuddhiststupa at Mihintale in 1990

XXX

CHAPTER

1

INTRODUCTION Robin Coningham

THE modem settlement of Anuradhapura is the capital of Sri Lanka's North Central Province and has a population of some fifty thousand. However, in the nineteenth century, it was little more than 'a small mean village, in the midst of a desert' (Davy 1821: 225) (Fig. 1). Anuradhapura's recent genesis owes much to the colonial restoration of the ancient irrigation system of the area (Parker 1909; Brohier 1934), but a great deal more to its role in the hearts and minds of many Sri Lankans as the royal capital of the island for over one and a half millennia. To some it is a symbol of the island's magnificent pre-colonial past, echoed on modem Sri Lankan banknotes and stamps; to others it is a holy city to be visited on pilgrimage; while to others still it is a reminder of Asia's rich cultural heritage, a heritage which has been barely exposed. It is a city which, through the pressures of modem politics, has been politicized from the massacres of 1986 to the restoration of the Mirisavatiya stupa by the late President Ranasinghe Premadasa in 1993. It is also a city which has kept the same name from its initial foundation in the first millennium BC to its selection in AD 1873 as the administrative centre of colonial Ceylon's North Central Province, despite abandonment from the eleventh century AD. It also represents, as noted by Anuradha Seneviratna, one of the world's major archaeological sites, covering over 40 square kilometres (Seneviratna 1994: 13). Although recognized and protected as a UNESCO World Heritage Site, the process of curation of these monuments and their environs was begun in the nineteenth century by colonial officials prior to the development and establishment of an Archaeological Survey in 1890. This process was greatly strengthened in the late 1950s when the Prime Minister, S.W.R.D. Bandaranaike, ordered the destruction of the structures of the administrative centre which had been established by the British in the heart of the ancient ruins of the capital of Anuradhapura and thus separated the city into two - a sacred city and a new town. While archaeological investigations continued at Anuradhapura under the auspices of the Archaeological Survey and the Anuradhapura Preservation Board, the scale and speed of research altered radically with the creation of the UNESCO - Sri Lanka Project of the Cultural Triangle by President J.R. Jayewardene in 1980. In Anuradhapura, the programme undertook to safeguard the Abhayagiri and Jetavana monastic complexes by excavating, conserving and presenting them to both pilgrims and tourists. At the same time the

Archaeological Survey continued to carry out research and in 1984 set up the Anuradhapura Citadel Archaeological Project (ACAP), under the direction of Dr Siran Deraniyagala, specifically to investigate the ancient urban core of the complex. This volume presents the results of a single trench, Anuradhapura Salgaha Watta 2 (ASW2), and associated fieldwork that was carried out under the auspices of the ACAP by a collaborative Sri Lankan-British team between 1989 and 1994 (Pl. Ia, Figs 2, 3). Archaeologically, Anuradhapura is an extremely important site as it fills a lacuna in the chronological and artefactual sequences for the island. Whilst major excavations have been conducted over the last thirty years at the major sites of Mantai, Kantarodai, Pomparipu and lbbankatuva, the associated reports are still only at a preliminary stage. This state of affairs has led to a reliance upon textual sources for much of the early history of the island and a subsequent relegation of archaeological research to support such sources. Such a process has not been without its problems (Coningham 1994a, 1995a), and it is hoped that this volume and its companion volume will illustrate some of the opportunities that archaeology can offer the historian. This process, combined with a general absence of chronometric dating, has led also to the lack of a classic type site for the island's chronologies. It is hoped that these two volumes, augmented by the future publication of the ACAP's sondages excavated between 1984 and 1990, will allow Anuradhapura to provide a typological artefactual and structural sequence with which to date other sites within the island. Anuradhapura also fills a lacuna in the chronological and artefactual sequences for the south.em part of South Asia. It can undoubtedly be classified as an Early Historic fortified city and, as the most southerly example of its type, helps us to understand the test models for this, the second urbanization of South Asia. Indeed, the nearest known examples of similar cities are at Dhanyakataka in Andhra Pradesh and Banavasi in Karnataka, some 900 km to the north. The presence of a fortified urban centre in the interior of the island in the fourth century BC surely recommends reexamination of earlier hypotheses that the urbanization of the peripheries of South Asia occurred as a direct result of Mauryan expansion and contact in the third century BC. Indeed, the excellent sequence of structures and artefacts at Anuradhapura allows us to study an aspect of the urbanization of South Asia in some detail, with development from a small Iron Age settlement to

Anuradhapura:The Site growing use of chronometric dating within Sri Lanka is helping to establish its position, not as a cultural cul-de-sac but as the pivotal point of South Asia. This is not to suggest, of course, that ritually Anuradhapura is not a site of more significance as a result of its association with Buddhism. This association has taken two forms, one in a physical sense of relics, the other in a more mystical sense, both of which are recorded in the Mahavamsaand Culavamsa,Pali texts which relate the history of the island. The first type of significance is given by the presence of relics associated with the Buddha, or Buddhism, within the monastic establishments in the city. The Mahavamsa records that many of these relics were taken to the island during the reign of King Devanampiya Tissa (r. 250-210 BC), shortly after his conversion to Buddhism by Asoka's son, the devout Mahinda. The relics taken to Anuradhapura included the Buddha's right collar-bone, which was enshrined in the Mahavihara's Thuparama stupa (Mvs.xvii.55-57); a branch of the Bodhi tree under which the Buddha had obtained enlightenment, which was enshrined in the Mahavihara's Bodhighara (Mvs.xix.35-46); the bowl relic, which was enshrined in the Cetiyapabbata close by at Mihintale (Mvs.xvii.22-24); and one of the original eight shares of the Buddha's remains from the stupa of the Koliyas of Ramagrama, which was then enshrined in the Mahathupa or Ruvanvalisaya (Mvs.xxxi.1-126). These relics of the Buddha were later supplemented by the arrival of the tooth relic from Kalinga during the reign of King Sirimeghavanna (r. AD 301-28). The latter was first housed in a building called the Dhammacakka within the Citadel itself (Cvs.xxxviii.92-98) before being installed in the Tooth Relic temple, or Daladage, during the reign of King Dhatusena (r. AD 455-73) (Cvs.xxxviii.70-72). This relic, a relative late-comer to Anuradhapura, was to become the symbol of kingship of the island and, when Anuradhapura became untenable, it was moved from capital to capital until in 1815 it was captured by the British. While Anuradhapura was thus associated with the Buddha through his relics, it was also associated with him in a more mystical way. The Mahavamsa records that Anuradhapura was founded as a village by Anuradha (Mvs.vii. 43), a minister of King Vijaya, who colonized the otherwise unoccupied island on the day of the Buddha's nirvana (Mvs.vii.1-4). The site was later settled by Prince Anuradha, brother-in-law of King Panduvasudeva, who built a tank and palace there (Mvs.ix.9-10). It was then selected by Prince Anuradha's great-nephew, King Pandukabhaya, as his new capital - Anuradhapura (Mvs.x.73-102). The association of the city with the Buddha is only later made in the Mahavamsawhen it describes the reign of Pandukabhaya's grandson, King Devanampiya Tissa, in the third century BC. Following the latter's conversion to Buddhism by Mahinda, he presented the royal garden known as the Mahameghavana, situated to the south of Anuradhapura, to the Sangha, or Buddhist order, and

a medieval metropolis. Unlike many of the great Early Historic cities in the north of the subcontinent, Anuradhapura is protected to ensure that it is relatively free of the pressures of increasing urbanism and agriculture. Another aspect of this importance is illustrated by Anuradhapura's pivotal role in Indian Ocean trade. Although rather better known sites such as Mantai (Carswell and Prickett 1984) or Arikamedu (Wheeler 1946; Casal 1949; Begley 1996) are frequently cited as providing clear evidence of the early and late stages of this trade, Anuradhapura, with its sequence from the beginning of the first millennium BC to the beginning of the first millennium AD, straddles its growth and development for almost two millennia. All the more surprising, then, that Anuradhapura is situated over 60 km from the coast, with no navigable river connecting the city to the sea. The second volume of the present report will provide ample evidence of this trade and contact with the coast in the form of early Islamic glass and glazed ceramics, Greco-Roman glass, metalwork and derived ceramic forms, Chinese glazed ceramics, imported semi-precious stones, as well as the presence of marine species at the site. Such studies have also allowed us to understand more about the position of Anuradhapura as a primate city within the island and its role as a centralized manufacturing centre. We have been able to study internal trade developments through analysis of metal-working, stone-working and shell-working debris at the site, allowing us to identify the stages at which different raw and semi-processed materials were processed within the site. A further, connected aspect is the evidence at Anuradhapura for the development of writing systems within South Asia. The earlier prophetic work of Deraniyagala at Anuradhapura suggested for the first time that Brahmi, the ancestor of many of South Asia's vernacular scripts, occurred a number of centuries earlier than had previously been thought (Deraniyagala 1990a). It had been generally accepted that this script derived from a Semitic script developed in northern India under the Buddhist emperor, Asoka, in the third century BC and had spread southwards through the peninsula until it reached Sri Lanka (Buhler 1896; Winternitz 1927; Dani 1963; von Hinuber 1990). Our own work now supports Deraniyagala's earlier hypothesis, and evidence of Brahmi script dating to the beginning of the fourth century BC is presented in Volume II. This discovery, the earliest example of its kind in South Asia, has enabled us to reassess the traditionally accepted theories and suggest fresh hypotheses for its development and spread through trade (Coningham et al. 1996). These combined archaeological factors help to overturn the cultural stereotype of Sri Lanka, which suggests that as it is situated at the southern tip of the peninsula it was the latest recipient of any innovation. As this theoretical paradigm appears to have been widely accepted, all resultant formulations of relative chronologies have naturally followed its directive. The

2

Introduction felt that the preparation of this report, six years after the final excavation season and four years after the last field study season, is not unduly excessive. The present volume - Volume I - contains a further six chapters which discuss the site, its location and chronology, and the excavations at trench ASW2. Volume II contains the artefactual chapters and has been divided largely according to the physical material (see p. 147 for a summary of Volume II's contents). In addition to the discussion of general aspects of each of these artefact categories, the actual catalogue data are also presented. While this block clearly makes up the largest section of Volume II, its position is key to the importance of the site. The publication of the artefacts from the excavations at trench ASW2 allows presentation of the data, enabling the conclusions presented here to be refuted or supported. It is hoped that their presence within Volume II will act as a catalyst. Our reasons for so doing are adequately summed up in Cunliffe's words (Cunliffe 1984: viii):

with Mahinda proceeded to mark out the future location of the various monastic monuments and structures. At each site Mahinda marked there was an earthquake and, on inquiring from the monk, Devanampiya Tissa was informed that similar establishments had been located in the same places during the lifetimes of the three Buddhas who had preceded the historical Buddha in the present era - Kakusandha, Konagamana and Kassapa. They had all visited the site in the past, when the city, the royal garden and even the island were known by different names. Thus the Buddha Kakusandha was given the garden Mahatittha by King Abhaya, when the city was known as Abhaya and the island as Ojadipa (Mvs.xv.56-59); the Buddha Konagamana was given the garden Mahanoma by King Samiddha, when the city was known as Vaddhamana and the island as Varadipa (Mvs.xv.91-93); and the Buddha Kassapa was given the garden Mahasagara by King Jayanta, when the city was known as Visala and the island as Mandadipa (Mvs.xv.125-127). The sanctity of Anuradhapura was further enhanced by a record in the Mahavamsa that Gautama Buddha himself, even before Vijaya's arrival, had visited the future site of the monastery and meditated at the future sites of the Bodhi tree, the Mahathupa and the Thuparama (Mvs.i.80-83). That these traditions were widely held is supported by the report of the fifth-century AD Chinese monk Faxian (or Fa Hsien), who visited Anuradhapura on a pilgrimage to the holy Buddhist sites and stated that the Abhayagiri stupa had been built over one of the footsteps of the Buddha, made when he visited the site (Beal 1869: 150). These factors, when combined with the series of monumental constructions erected by successive kings at the city, continued to add to the ritual significance of the site to Buddhists. Indeed, Anuradhapura contains seven of the island's sixteen holiest places of Buddhist pilgrimage: the Bodhi tree, the Maricavatticetiya or Mirisavati stupa, the Mahathupa or Ruvanvalisaya stupa, the Thuparama, the Abhayagiri vihara, the Jetavana vihara and the Selacetiya (Geiger 1960: 207). Even in the late eighteenth century AD this factor still caused Sri Lankan monarchs, by then confined by the European maritime powers to kingdoms in the hill country, to undertake pilgrimages to the ruins of the holy city (Coningham 1994a: 92). Anuradhapura is therefore not merely a ruined city, but a living cultural entity which intertwines identity, ritual, tradition and archaeology. The report of the Sri Lankan-British excavations at trench ASW2 has been divided into two volumes Volume I: The Site and Volume II: The Artefacts. It is

... no excavation report, however detailed, can hope to be more than an interim summary of a site. To suggest more would be naive or arrogant. A data-set of this kind . . . will continue to be reworked by students for the foreseeable future asking new and increasingly sophisticated questions. These reports merely advertise what is available and offer some general approximations to the truth which may help those interested in these matters to design new and more penetrating analyses. Before commencing, certain conventions adopted in the text should be explained. Firstly, De Silva's list of Sri Lanka's rulers (see Appendix A) has been accepted as an initial framework for the island's chronology (De Silva 1981). We understand fully, however, that as this was based upon a combination of sources, including the Mahavamsa,the Culavamsaand various inscriptions, it is not necessarily free from error or omission (Coningham 1994a, 1995a). Secondly, for the sake of consistency, Bandaranayake's terminology and names for the monuments within Anuradhapura have been adopted within the text and illustrations (Bandaranayake 1974). Finally, it should be noted that all diacritical marks have been dispensed with, following the convention used in the Cambridge Encyclopedia of India, Pakistan, Bangladeshand Sri Lanka. (Robinson 1989).

3

NORTH I

-,

'

,I

'

'

'

I

I

I

/

I

/.f;

I

,_

I

.,I I'

I

'

'-'

//

,'

\_

-,

)

✓,

,

RUSSIA

I

'---L----------~-

1

'

'\. I

:\_

'

I

I \

\

\

le I I

I'

•-;;,,,_.,, $

~,,,

,1

'

\

\

' ' \_

N- AFRICA

'\f,:I}

-0

--..,_.

1,1•~·'"-

'\:'..:'.6_ ... ,·.~

,,,.,

'

~ '

Cairo

~.,:i.:~

- ' -- _

•• ,.·

\).

I

•

'

'

;',

✓

,_.,,,.'

CHINA

,.(.

'

!:··

I

\

.,

·:-\ / • r PERSIA•

.Kabul

edelhi \

-

I

\ \

' ',

\

11

'

\

KEY

........

·'. I

GREAT WALL OF CHINA ~,' SILK ROUTE

--~~.!':::,·, .... -. '\

- --

.,.

-~··

....~

\

I

,_

...

--------__

,,EGYPT

\

I

JQ~::.-.-: _.-:,;,·,:·· ,.-~-:----'.-.-::-:.r-----/._/ ' '"·· .. A ~ ._····\r.._~. _/ ,:::_::_--·":' ?;,-,cc,,;{._: :.,/' \\ . i;~r\:\: t

:(\1

{\:.:, I

If"""',• -----,--~ \ ,-\') I

!

-:.:.:~' ..

V'I

I

\

'·

'CL:.:t•...:

'

I I

GRID AT 20 DEGREE INTERVALS - - I- -

\

Indian Ocean\

Fig. 1: Map showinglocation of Sri Lanka and Anuradhapura

... ~ S' ~

z

:c ,.., :,0

C

ii c.,

~

:x

_J[ CITADEL IC


V,

....

-

:,:, "' ,.., V,

j[QJI ,.

---

r-

= I ClTAOEl I=

"'

Cl

,. GI

Sl ...

>

0

;: :x Cl z

r-

>

"'

"' XI

0

.... "' ;:;; V,

..,


z:,,:

:,0

> z :,:

"' :x "' ~ c;\

"' HONASTEIUES

VILLAGE AMII TANK

FOREST ANO HfRHITA6E

Fig. 7: Schematicplan of Anuradhapura (after Seneviratna 1994)

30

Thec·ity of Anuradhapura

2m

0

AG-69 (afterDeramyagala . Fig · 8·· Sectwn . of sondage . 1972)

31

Anuradhapura: The Site

N

t

..___,

Key

[QBSurvey ■

Area

Areas covered

■

A-D by ASW2 Environs

Plan

•

Excavation Pillared

Halls

Fig. 9: Plan of the Citadelshowingthe locationof sondages

32

Trenches

The City of Anuradhapura

Fig. JO:The westerndefencesof the Citadelshowingrampartson right and silted moat in foreground

33

Anuradhapura:The Site

!I,,~,

G ~~

-PLAN·-

OF" HOUSE.S-

IM THE CITADEL

I

l 6 metres

Fig. 11: Plan offinal occupationphase street (afterAyrton 1924)

34

The Cityof Anuradhapura

L

PLAN or

CITADEL

ANURADHAPURA

\

\ i

\

\

II II

""" ""n ,, fl ,, " ,: .II ",, "" "",, ",, ,, II·

\ \

II

\

II

,, ,, ,, ,,. ,, ,, ,, ,, ,, ,, ,, II ,,,, ,,

.."

·:

II

., ,, \\

\\\ \

'\'\

'\

,,,,

(,),, /j

,,

\,

\.. \

,::.:

I r--..:.-==:.:::~:: I ',._

.,

/

ti

I

::

...

1,

i!.

", ., \,

,,JI

.ii ,' l:J

., ., :,

--... II,,,

il

~·. ..

It

if

~

.

. ::,

if

•/

!:

•r

==::----....... -~,,:

lj

II

it ,,,,

\•

,,,, I,

, I

,/ ,,' I I

I

0

I

I

I 100 metres I

-

Fig. 12: Plan of Paranavitana'sexcavations(afterParanavitana1936)

35

Anuradhapura:The Site

Fig. 13: Building A

Fig. 14: The Gedige

36

The City of Anuradhapura

Fig. 15: Augeringin front of 'Vijayabahu'spalace'

37

Anuradhapura: The Site 0 N II

• • • •

a

•

II"" 0

It) ,..

Cl)

e

a:

•

0 ,..

w 1w

::E It)

Q

a:

~ l-

0

a: ::) 0

(.)

a: w 1::)

0

·.: . : .·..

·:. ... . . ·... ·.. : ...: .. ·... . ..··.. . ·.....

:·:. :·-:, ....... . :·,. . . .-•,·: .·.: .... . ·: .·. ·_. ' .. : . ,· ..... '··.· ···....·.·

Q

'•

a:

~ t-

a: ::)

... ·.· .. ·.· .. :·. .. : .

. :.

0

;:·..:.·. ::

(.)

...

.... .. .

.....

a: w

w

z

....

~

·..·.

·'

..... . -·....... .

•

• •

• ~.• ::-t"r-.=---=.--1 •

.. . ..

. ·.. . . ... . .

•

::)

·.. . .:. ...:-

>

.. . ...

.. . ... .. . : .. . ..... ........ . . . ·.· .... ...

•

1

•

;

.. ..·. ·...

'·

·...\ : . ......... • .. ,'••

.•

·. ·..

i. •

:

r,:;=:;::::::=:;======:..l:· .....·..

·..

.. . ..

. . . .·.· . ... : ·_.. ·_ :.·.

38

... · .·........:• . . . ....: .. ·.:...: ; ·.-........ . . ... _

< ...J < C.

...... . . ·.. ... ·... ,·.. ·.....·.. . ... ·. ...: -:. ..... : ...

•

•

CJ

:r

< al

~

< ..,

The Cityof Anuradhapura

Fig. 17: TheBodhi tree shrine

Fig. 18: The Thuparamastupa 39

Anuradhapura:TheSite

Fig. 19: TheMirisavatistupa

Fig. 20: The Jetavanastupa 40

The City of Anuradhapura

. Fig. 21: TheBasavakkulam

Fig. 22: The Tissavava

41

Anuradhapura:The Site

Fig. 23: The Nuvaravava

Fig. 24: The Toluvilastupa

42

The City of Anuradhapura

Fig. 25: The Westernmonasteries

Fig. 26: The Vessagirivihara

43

Anuradhapura:TheSite

Fig. 27: An earlyBrahmiinscriptionat Vessagiri

Fig. 28: TheMeghagirivihara

44

The City of Anuradhapura

· Fig. 29: Thenorthernannicuton theMalvatuOyain theprocessof beingencasedwithina modemdam

Fig. 30: TheKalavava

45

CHAPTER4

THE FORTIFICATIONS Robin Coninghamand Paul Cheetham

4.1 Introduction The fortifications or defences of the Citadel were first identified on the southern side as a banked earthwork by Parker in the early twentieth century (Parker 1909: 274), and by 1924 the entire circumference was included in a map of Anuradhapura published by the Archaeological Survey Department (Hocart 1924) (Fig. 31). They were excavated for the first time in 1960 by Godakumbura, who cut a section through the southern rampart (Godakumbura 1961). He hypothesized that they had consisted of a moat and a brick and earth rampart capped with an ashlar wall. Godakumbura's successor, Silva, continued this work by excavating part of the eastern rampart in the 1970s. Although the report is still pending, it appears that he uncovered a brick wall 5 m high. In 1992 the fortifications were further excavated by a team of Japanese and Sri Lankan archaeologists, who may have identified a defensive sequence of over 1500 years (Coningham 1993; Ueyama and Nosaki 1993). These studies suggest that Anuradhapura, like many other early historic cities, had a defensive complex consisting of a moat and a rampart capped by a wall. In 1993 and 1994 the British sub-project working within the Anuradhapura Citadel Archaeological Project (ACAP) decided to study the course and fabric of the Citadel's fortifications. Our survey strategy had two aims: firstly, to identify the wall and moat; and, secondly, to evaluate the possible advantages of using archaeological geophysical field techniques in Sri Lanka. The reasons for these aims were threefold. Firstly, we wanted to test the feasibility of using a proton magnetometer in Sri Lanka. As we were aware that a single-sensor proton magnetometer would be unsuitable for detailed work owing to a combination of factors, a gradiometer arrangement normally being more appropriate for small-scale anomalies as well as reducing interference effects such as diurnal variations in the earth's magnetic field (Milsom 1989: 45; Clark 1990: 66; Scollar et al. 1990: 455), we decided that for the 1993 field season we would attempt to record a feature which should be identifiable as a clear anomaly - the moat. Our survey sectors were therefore selected to cross the postulated line of the fortifications at right angles to allow its identification. Secondly, following our involvement in the interpretation of the excavation section across the southern rampart in 1992, we intended to identify the course of the wall around the entire site in order to

draw attention to its course for preservation and protection from either housing development or use as a quarry for building materials, as is currently the case for the southern rampart. Thirdly, we wanted to plan the course of the wall in order to estimate the extent of the area it enclosed. This chapter is divided into five sections. The first covers previous investigations of the Citadel's fortifications, the second describes the various techniques and methods used to investigate their course, and the third and fourth describe the investigations on the north, south, east and west sectors of the site. The fifth and final section introduces evidence for dating the defensive complex of the Citadel.

4.2 Previous investigations of the Citadel's defences Although textual descriptions of the city had been available from 1837, following the publication of the translation by George Tumour ( 1779-1843) of the Pali text of the Mahavamsa,the secular urban core of the monastic complex of Anuradhapura - the Citadel - was not successfully identified until the early twentieth century, when Parker traced the southern wall through a combination of surface observations and the Mahavamsa's historical topography (Parker 1909: 274). Surprisingly, it was not until 1960 that the identified rampart was first investigated, despite the fact that it had been surveyed and incorporated into one of the first detailed Archaeological Survey Department's plans of the area (Hocart 1924). The then Archaeological Commissioner, Dr Godakumbura, initiated an excavation on recently acquired land which incorporated one of the southern gateways and part of the southern rampart. One north-south trench was cut between the Godage private road, which runs parallel to the rampart within the walls, and the irrigation channel to the south of the rampart (Godakumbura 1961). The 5.66 m deep trench located two parallel lines of ashlar blocks some 3.5 m apart, packed with filled earth and occasional stone slabs. Godakumbura proposed that the modem irrigation channel running parallel to the rampart probably followed the line of the old moat. He also appears to have assumed that the rampart was a single-phase construction and suggested that the parallel lines of ashlar slabs lay at the centre to give the construction strength while the base was probably of brick. A larger east-west trench was cut in order to expose a portion of gate and a 17 m wide street passing through it at right angles to the rampart. At the centre of the gate a brick structure measuring 8 x 5.33 m was identified and has been interpreted as a check-post. The excavation was never fully completed or published, and no attempt was made to date,

Anuradhapura: The Site

phase, record or conserve the structures uncovered. The unprovenanced finds stress the mixed nature of the deposit and include sherds of Rouletted ware, Sassanian-Islamic blue glazed ware and early Islamic white glazed ware or early South Chinese white glazed ware (ibid.). A large section on the eastern side of the Citadel was excavated by Godakumbura' s successor, Dr R.H. de Silva, in the 1970s. Although the structures exposed were conserved, the excavation report is still pending; however preliminary results suggest a wall some 5 m thick and 5 m high with a 2.5 m parapet on the inside (Pl. IVb). The foundations were dated to the second century BC. In 1992 the southern fortifications were further investigated by another ACAP sub-project, directed by a team from the Japanese Overseas Co-operation Volunteers (JOCV) (Coningham 1993; Ueyama and Nosaki 1993). This team selected for excavation a very well preserved section of wall which we had surveyed in 1991 and 1992, close to the Sanghamitta Road which cuts north-south through the Citadel. In addition to clearing a stretch of 20 m along both sides of the wall, later conserved for public presentation, they excavated three sondages, pit A, pit B and pit C, down to bedrock (Ueyama and Nosaki 1993). Although no samples were selected for chronometric dating, the excavators suggested a date from the late Anuradhapura period for the final phase of the fortifications (ibid.: 99). The results of this excavation are re-examined in more depth in section 4.8 below. Having thus summarized the history of archaeological investigation of the Citadel's fortifications, we may now detail the results of our own two field seasons of work on them.

4.3 Methods and techniques As part of the contour survey of the Citadel, members of the Archaeological Survey Department and British-Sri Lankan Anuradhapura Project carried out a detailed clearing operation and survey of the Citadel from 1989 onwards. Part of their brief was to trace and plan the extent of the site itself. During this work it became clear that information concerning the nature and orientation of the city's ancient fortifications needed to be recorded. One of the clearest alignments noted was a parallel line of ashlar blocks running eastwards from the Sanghamitta Mawatha bridge across the old southern moat to the southeast of the Godage Walawwa and the location recorded as a prospective excavation site. This site was later excavated by a Japanese sub-project in 1992 (Coningham 1993; Ueyama and Nosaki 1993). It was then decided to investigate the defensive circuit by the use of four methods: surface survey, magnetic survey, resistance survey and a coring survey. In particular we wished to assess the possible contribution of archaeological geophysics to the archaeology of South Asia through a pilot study at Anuradhapura. These techniques, which are virtually non-destructive, now play an integral role in archaeological site assessment

in Europe. They are generally used in two main applications: firstly, the assessment of newly discovered archaeological sites and, secondly, the comprehensive survey of a known site. While the former usually aims at defining the extent of the sub-surface remains, the latter can be used as a base for the creation of a site management strategy to assist with any future development at the site. These techniques have a further advantage over test excavations or sondages: they are highly cost- and timeefficient, so a small team can survey sites quickly. Many of these points have been fully illustrated by two UNESCO pilot missions to South Asia led by Coningham in 1997. These missions, one to Bangladesh and one to Nepal, attest to the effectiveness of this instrumentation in detecting subsurface monuments within the context of archaeological sites (Coningham and Schmidt 1997a, 1997b). 4.3.1 Surface survey At each sector we constructed a grid using the concrete posts which had been erected at 30 m intervals across the entire archaeological site. As all our sectors were located in dense scrub they had to be cleared, and while doing this we uncovered a number of ashlar blocks and scatters of brickbats. The surface survey teams planned the visible ashlar and brick debris and alignments and conducted a detailed contouring survey with a theodolite every 2.5 m within the respective sectors. In a number of sectors, the north, south and west in particular, the final phase of the Citadel's fortifications - an ashlar wall - were clearly visible. This surface observation later proved to be very useful when attempting to interpret the archaeological geophysical results. 4.3.2 Proton magnetometer survey We constructed a grid at each sector using the concrete posts which had been erected at 30 m intervals across the entire archaeological site. After the sectors had been cleared of dense scrub, we conducted a geophysical survey with a proton magnetometer. The survey was conducted with the single-sensor Geometrics Portable Proton Magnetometer, model 0816. The model selected had the advantages of providing rapid, accurate measurements (one reading every six seconds) while being a rugged and compact field instrument. It was powered by twelve 1.5 volt batteries and weighed some 4.3 kg. The data from the six survey traverses that made up each of the transects were analysed using Spyglass Transform for Windows. The six traverse readings were taken across the transect as the survey progressed along the transect length. As a single-sensor instrument was used, any anomalies of archaeological origin would be superimposed upon the diurnal variation, this variation being evident as the general rise and fall in the profile along the transect. In each case the transect's magnetic profile, magnetic contouring, magnetic dot density image (higher readings showing as darker areas) and topographical profile have been illustrated. 4.3.3 Earth resistivity survey We returned to the Citadel in 1994 in order to further clarify the investigations with an earth resistivity survey. The

48

TheFonifications electrical resistivity of the earth largely depends on its moisture content, which varies between differing subsoil features. Such sub-soil features, whether human or natural, may appear as anomalies if their moisture content differs significantly from that of adjacent features (Clark 1990: 27). It is possible to detect such anomalies by measuring the varying resistance to the passing of an electric current through the soil between two probes. Although this is an over-simplification, the design of all resistivity meters is based on this model. In practical field instruments four probes are employed: two to pass the current through the ground and two to measure the resistance. Frequently, but not under all conditions, features such as moist ditch fills will register lower resistance readings, while others, such as sub-soil walls, will register higher resistance readings, when compared to the mean background level of resistance. The method is affected by climatic and geological conditions, which can significantly enhance or mask traces of human activities (ibid.: 53). Because of increased numbers in the 1994 field teams we were able to concentrate on area surveys rather than linear ones. The area survey is now the norm for resistance survey as it allows identification of possible man-made features with a greater degree of certainty. The equipment used was the Geoscan Research Resistance Meter RM4, which is designed to be a swift, robust and accurate field instrument. The battery had a life of 22 hours and took 14 hours to recharge. For an archaeological survey it is normally used in the 0.5 m twin probe configuration. This configuration provides simple response profiles, has good depth penetration, is not affected by probe orientation and is efficient in use (only two probes fixed on a rigid frame are moved between readings). The spatial resolution of the surveys was 1 x 1 m, giving 400 readings per 20 x 20 m surveying grid. The data was hand-logged on prepared survey sheets and the readings were analysed in the field with Geoplot software. Contors software, written by J.G.B. Haigh (University of Bradford), was employed for more detailed analysis and the creation of the resistivity survey images. Further data analysis and the creation of the profiles was done using software written by P.N. Cheetham (University of Bradford). Micrografx Windows Draw 3.0 was used to create the final publication figures. 4.3.4 Soil auger coring survey Soil auger coring has long been practised in the Netherlands for building up compilation maps of soil types (Steur 1961). Attempts have even been made at conducting close-interval coring of archaeological sites in order to predict detailed internal site structure (Hoffman 1993). However, the best results are obtained when dealing with sites at a macro-level (van Andel and Runnels 1995). In 1994 we also wished to test the interpretations of the geophysical and surface surveys with a soil auger. Using a collapsible, 10 m long Eijkelkamp soil auger

for heterogeneous soils, kindly lent by the McDonald Institute for Archaeological Research (Cambridge University), we took cores to bedrock along transects at right angles through the defences and out into the surrounding fields. Each 0.2 m soil core was recorded for Munsell colour, texture and inclusions. Using the results of these cores we were able to reconstruct the stratigraphy from these samples and draw a section allowing us to confirm or refute the presence of the defensive ditch. Drawbacks of this method include the inability of heads to grind through stone, although potential damage to a site or objects is statistically minimal. It is a rapid and cheap method of sub-surface investigation. We found that we could complete a 10 m deep core in about 4.5 hours. We also successfully conducted a complete core profile across the Citadel mound (Fig. 32). These cores, taken at 150 m intervals, allowed us to build up a projected macrostratigraphic profile for the entire site as well as confirming that the earliest occupation at the site was on a slight rise of alluvial gravels and bedrock. We have since conducted a similar survey at the Bala Hisar of Charsadda during our collaborative fieldwork with the University of Peshawar and again found very positive results (Ali et al. 1998).

4.4 The northern fortifications The northern edge of the Citadel mound was surveyed in both 1993 and 1994. The 1993 survey area measured 100 m north-south and 30 m east-west (Fig. 33). The profile of the northern edge of the Citadel mound is similar to that of the western edge. The crest of the rampart stands some 5 m above the surrounding paddy fields. The paddy begins at between 71 and 78 m from the beginning of the transect and gradually rises in gradient at 100 m. Once the bush had been cleared it became evident that there were numerous scatters of ashlar blocks on the slope of the rampart. In addition, it appeared that part of the ashlar wall identified at the southern and western edges of the mound was also preserved in situ at the northern edge (Fig. 34). A number of slabs orientated east-west formed an alignment 15 m long. This alignment crossed both transects at between 20 m and 25 m from the beginning of the transect. Two transects, each 100 m long and 5 m wide (six traverses spaced 1 m apart), were surveyed using the proton magnetometer. Transect 3 indicates a high degree of magnetic variation along its length both within the Citadel and down the slope onto the paddy (Coningham 1992). A significant band of negative values across all six traverses is noted at 40 m, which could indicate the line of a stone feature. Transect 4 is magnetically less active although still exhibiting some anomalous variations at around 40 m. Both transects peak at around 50-60 m with evidence of larger-scale variations within this band, and both then drop down towards the 80 m point before rising again. These profiles suggest that this may represent changes in sub-surface deposits and not simply a diurnal variation effect. In 1994 we conducted a resistivity area survey covering 4200 square metres. As expected, the Citadel mound itself gave readings of highest resistance (measuring up to 37 ohms), probably the result of a combination of the dry49

Anuradhapura: The Site season climate and the fact that the man-made tell site stands some 4 m above the surrounding paddy fields. This area of highest resistance does show some evidence of rectilinear edges to anomalies that may represent structural features (Fig. 35). The ashlar slab alignment identified in the 1993 surface survey failed to register on the resistivity survey. Similarly, the only feature within the paddy fields was a higher resistance along the paddy bunds on either side of the small stream in the northeast corner of the survey area 'C'. During the 1994 season we also took six auger cores along a 150 m transect running at right angles across the defences and out into the paddy fields (Fig. 36). During the magnetometry and resistivity surveys we had failed to differentiate any clear anomalies outside the rampart, although the magnetometry survey showed some anomalous responses. The auger coring, however, allowed us to identify a feature, probably connected with the Citadel's fortifications. We first plotted the profile of the surface and then the depth and contouring of the bedrock below, as indicated from the cores. It was clear that the bedrock was 1-2 m higher to the south, that is within the Citadel, as opposed to that underlying the paddy fields. It was also clear that the bedrock had been cut to a maximum depth of 2 m by a feature at a distance of between 45 and 90 m along our transect 'D'. This feature had in turn been cut by a feature filled by a silt rich in snail shells, 'E'. The latter feature was 80 m wide and some 3 m deep. It seems possible that the rock-cut ditch is part of the original fortifications, while the shell-filled ditch is a laterintrusive phase of fortification.

4.5 The eastern fortifications In 1993 we surveyed a block 30 m wide and 65 m long with an additional 10 x 10 m block at its extreme southeast corner (Fig. 37). The eastern sector represented both the easiest and the most difficult area to work in. Although we were able to clear collapsed material and bush from an old excavation trench and expose a 4 m high brickbat wall running north-south, the gradient recorded by the contour survey showed a very gradual profile (Fig. 38). It also appears that the road on the eastern edge of Sector C had been built over part of the rampart. On clearance we found few ashlar blocks, and none in situ. As we were also unclear as to the course of the rampart at the southeast of the Citadel we conducted a surface survey of a further area to the east of the main block of C. When this fresh area, measuring 90 m north-south and 30 m east-west, was cleared, we found a low bank with a north-south alignment of blocks which have been interpreted as marking the course of the wall. Three magnetometer transects were taken over the sector; only Transect 6 is described (Coningham 1992). Transect 6 (k-1), measuring 65 x 5 m (six traverses spaced 1 m apart), was located 5 m to the north of Transect 5. The readings show relatively small

variations from Oto 30 m, with a negative dip at around 18 m that could indicate a buried stonework feature. Between 30 m and 57 m stronger, larger-scale variations are apparent in the profile. A steeper drop in background levels beyond 57 m is more than may be expected to result from diurnal variations and so may represent a general change in the subsurface deposits. In 1994 we investigated the eastern sector using an area resistivity survey covering 2100 square metres. The area survey, together with a resistance profile along the line of the auger transect (see below), is reproduced in Figure 39. This, one of our most successful surveys, indicated significant, substantial sub-soil features. The most obvious feature was the 30 m long high-resistance linear anomaly, aligned north-south, between 20 and 30 m east of the sector's western edge (marked 'A - A'). This anomaly, notwithstanding what is interpreted as a large robber pit (low-resistance anomaly 'Al', which lies on the line of the resistance profile), correlates with a number of ashlar slabs lying on the surface and most probably indicates the course of the wall. It is possible that a parallel north-south concentration of high resistance some 4-5 m further west may present another line of ashlar and brick. Such a wall would be very similar to the exposed portion visible at the southern sector. From this point to the western edge of the survey some structure is evident in and between a number of high-resistance anomalies that may represent building remains. Two of these anomalies are crossed by the resistance profile. It is very possible that low-resistance anomalies'D' represent the silt-filledcraters of robber pits or areas free from substantial building debris; anomaly 'D1' is substantially lower in resistance than the surrounding areas. An almost 20 m wide north-southalignment of low resistance, noted as feature 'C - C' on the figure, is interpreted as a possible moat or ditch, its edges being well defined on the resistance profile. To the southern edge of the survey area this anomaly is less well delineated, but analysis of the resistance readings in this area suggests that fans of higher-resistance material may have been dumped or slumpedinto the ditch at this point both from the west and south. A further area of lower resistance (markedly uniform), 'B - B', was identified at the easternmost edge of the sector, separated from 'C - C' by a 7 m wide, north-south aligned band of comparatively higher resistance, 'E - E', that also shows up well on the profile. It is significant to note that the feature A'-A', the fortification wall, runs due north-south across the survey grid, indicating that the line of the defences follows the 75 m contour at this point. This categorically refutes, for the first time, the postulated line of the eastern fortifications as indicated by Hocart, who suggested that the fortifications followed the 81 m contour, thus giving the Citadel a pentagonal shape (Hocart 1924). It now seems more likely that the Citadel was originally laid out as a rough square, but that the southeast corner has been badly damaged by erosion and agriculture. Only further detailed survey in this area is likely to confirm the exact course of the fortifications. The eastern sector was subjected to an auger coring transect measuring 140 m. A total of ten cores were taken

50

The Fonifications along this length and the stratigraphic profile was plotted (Fig.40). The resultant profile identified three very clear features which related to the results of both geophysical surveys. We first identified a 2 m deep and 70 m wide cut into the bedrock under the present edge of the Citadel mound 'M'. It is assumed that this represents an old moat or ditch, and it was identified as low-resistance anomaly 'C - C' on the resistivity survey and as a more active portion of the magnetometry profile. This feature was then in tum cut by a new ditch or moat, 'N', measuring some 65 m wide and 3 m deep. This second feature was filled with silts and snail shells, suggesting the presence of slow-moving or still water. This feature is clearly the low-resistance anomaly 'B - B' and is possibly represented by the magnetic response change beyond 57 m. That the two features were not contemporary is confirmed by the absence of shells in the inner, older moat and the clear shell horizon fill overlapping the inner moat. Anomaly 'E - E', a band of high resistance, was identified during the auger coring as a residual stump of bedrock isolated between the two moat cuts. It is clear that these two cut features are very similar to the double cut feature identified in the northern transect. It seems probable that the shellfilled features are later fortifications replacing the earlier, presumably silted moats.

4.6 The southernfortifications A brief surface survey of the southern edge had been made in 1991 by members of the Archaeological Survey Department and British sub-project and had been followed up by excavations in 1992 by members of the Archaeological Survey Department and Japanese Overseas Co-operation Volunteers (Coningham 1993; Ueyama and Nosaki 1993). However, this area was not suitable for us to examine in 1993 since the wire fencing or power lines that run along the entire length would have distorted the proton magnetometer survey. In 1994, however, we were able to examine this area during the resistivity and soil coring surveys (Fig. 41). The southern sector represented our largest area survey covering some 7400 square metres, running from the edge of the Citadel mound down to the Thuparama and Sanghamitta stupa complex to its south (Fig. 42). The results were remarkably successful, but only in combination with the results of the auger coring survey. Our survey sector was disturbed by a number of modem features, the irrigation ditch, the road, the foundations of buildings close to the cross-roads 'O', and grit thrown up from the excavation of the ditch 'P'. Other archaeological features included the southern extent of the rampart and fortifications 'Q' and a 30 m wide, east-west anomaly of high resistance, 'R', running the entire length of the survey area. This feature masks, or partially masks, areas of low residence 'S' - areas which we know are linked from the auger coring survey. One such area appears to have been delineated by lines of high resistance,

probably walls 'T - T'. Many of the high- or low-resistance anomalies identified during the resistivity survey were confirmed as archaeological features during the auger survey (Fig. 43). We took over a total of 18 cores along a length of 130 m from within the Citadel mound to the Thuparama complex. Initially we took one core every 2 m; however in certain localities along the north-south transect we cored more intensely. The base of the moat feature, for example, was sampled by no fewer than eight cores. While constructing the overall transect section we plotted only selected macrofeatures. Initially we plotted the ground surface and then the surface of the bedrock. It became clear that the surface of the bedrock sloped from north to south. The bedrock under the Citadel's rampart was some 2 m higher than that close to the Thuparama. It was also clear that there was no gentle gradient between the two ends of the transect; rather a large, scooped 'U' measuring 60 m wide and 3 m deep had been cut into the bedrock. This feature corresponds with the lowresistance anomaly 'S' identified during the resistivity survey. The bottom 1 m of this feature was filled with silts and snail shells - suggesting the presence of slow-moving water. This feature is clearly a silted moat. Owing to the paucity of finds from the cores, we are unclear as to the age of the cutting of the structure, however we are more clear about its abandonment. Feature 'R' was identified as a 6 m wide and 1 m deep deposit of grit and brickbat fragments running east-west across the transect. The surface survey identified numerous ashlar pillars and blocks within the feature. The similarity between this feature and others dating to the later phases of occupation within the Citadel, in particular a pillared alignment parallel to the Vijayabahu palace site, suggests a late Anuradhapura-period date for this feature. As it is clear from the coring that feature 'S' seals the silted moat or ditch below it, this suggests that by the late Anuradhapura period maintenance and use of the southern ditch or moat had already lapsed. It is interesting to note that the moat fill is rich in snail shells. Such a feature makes it tempting to link it with the shell-filled moat features at the northern and eastern sectors. It may be postulated that the fortifications were not relocated on the southern side of the Citadel because of the closeness of the religious structures to its south. In such a case it may be that the southern moat or ditch was just re-cut on the same alignment (Fig. 44).

4. 7 The western fortifications The area surveyed on the western edge of the Citadel in 1993 measured 107 x 30 m (Fig. 45). This area is one of the clearest, with the rampart rising almost 7 m above the paddy. The rampart stops at the edge of the paddy, which is some 54 m wide (Fig. 46). At the extreme western side of the paddy, 127 m from the eastern edge of the transect, the land rises slightly to the road. During 1991 this area had been selected for possible investigation as there were a number of ashlar blocks on the 82 m contour line at the top of the rampart mound. This alignment runs for some 180 m on a north-south alignment. On clearing the area for survey it became clear that the alignment consisted of two lines of parallel ashlar blocks spaced some 4 m apart. The alignment 51

Anuradhapura:The Site crosses the first transect between 25 m and 35 m from the beginning of the transect. It is very probable that this represents a continuation of the wall identified in the sixth phase of the southern rampart excavation (Coningham 1993: 114). Although surveyed with the proton magnetometer in 1992 (Coningham 1992), the area was flooded for paddy in 1994, causing us to concentrate our resistivity and auger coring surveys on the northern, eastern and southern sectors. Two transects were recorded in 1993 using the proton magnetometer, the first 107 m long and 5 m wide, the second 124 m long and 5 m wide, the latter also including a stretch of 20 m east into the Citadel proper. As with Sectors Band C, each transect was 5 m wide, and thus for each metre through the defences at right angles we recorded six readings. Transect 1 (a-b) was begun on the Citadel. Between 5 m and 15 m the magnetometer registered an acute positive anomaly larger than 160 gamma. This anomaly has been marked 'V'. As this anomaly occurred on the line of a wire fence that we had removed from a line of bushes, it is hypothesized that it represents nails or rusted metal fragments left behind. After this anomaly the readings display only minor variations until we reached a point at 50 m along the transect. Between 55 m and 65 m there are stronger variations with indications of a large positive anomaly. Both sides of this anomaly have been marked 'W'. The readings then exhibit only minor variations until 80 m is reached, when the variations become more marked. At 100 m we reached a house compound and the electrical disturbance produced a further acute magnetic anomaly which has been marked 'X'. Transect 2 (b-c) was not as revealing as Transect 1. It measured 124 m by 5 m and started 20 m further into the Citadel than Transect 1. Between O and 20 m we recorded a strongly variable set of readings. Between 20 m and 35 m we recorded an acute positive anomaly with a negative spike at its centre. This anomaly has been marked 'Y'. As with Transect 1 this is likely to be metal from the fence line running along the top of the rampart. The readings then show variation increasing up to 110 m. At 120 m a house compound was reached and a further positive anomaly was recorded. It is likely that the paddy fields at the foot of the rampart actually represent a silted portion of the original city moat, however only an auger profile will confirm this hypothesis.

4.8 Dating the fortifications Following the findings of the surface survey, part of the southern rampart was identified as being wellpreserved and suitable for an excavation to recover evidence of the dating and phasing of the fortification complex at Anuradhapura. Thus an ACAP team, in collaboration with a Japanese group, began excavation in 1992 with the aim of producing a datable sequence of construction for the Citadel's defences and to conserve and clear a length of rampart for presentation

to the public (Coningham 1993; Ueyama and Nosaki 1993). The coordinates of the deep sounding trenches, Anuradhapura Citadel Rampart South (ACRS) 5A, 4A and 4B, were 17N/16E. The dimensions of the three cardinally oriented trenches were 3 m long, 3 m wide and 8.1 m deep. ACRS 5A was located on the north edge of an observable parallel line of ashlar slabs, ACRS 4A was located on the southern edge of the slabs 3.5 m due south, with ACRS4B a further 3 m due south. The coordinates of the presentation trench were 17N/16E. Its dimensions were 24.5 m long, 15.5 m wide and 2 m deep. The trench ran along the line of the wall from the eastern side of the Sanghamitta Mawatha, or road, at the latter's bridge across the old southern moat to the east. As this excavation represents the first published report of a section through the ramparts of the Citadel, it will be examined in detail in this section since it offers an opportunity to date the successive phases of fortifications at the site. This section has been augmented with data recovered by an auger survey conducted at the site by our team in 1994 (Fig. 47). The individual contexts identified in the three excavation pits may be grouped to form a continuum of eight macrocontexts. They are, in order of age: Reddish Brown Earth, a mixture of Reddish Brown Earth and bedrock, a mixture of clay and sand lenses, an ashy-silt deposit, a decayed brickbat deposit, a further ashy-silt deposit, a gritty deposit and a topsoil humus. These eight macro-contexts are known to form the complete depositional sequence throughout the Citadel mound, and thus we are able to correlate the phases of rampart construction with a particular phased development sequence. During analysis of the data and sections it became clear that the three trenches, originally orientatedto excavate the rampartnear the latest ashlar phase, had only located the innertoe of the earliest rampart phases. The rampart centre had moved some 10 m north over time, so the description of the earliest phases of its construction is incomplete. The earliest phase occurs only in pit ACRS 4B, where the inner toe of a mound of compacted Reddish Brown Earth was encountered. The visible dimensions were at least 2 m wide and 2.10 m high, with redeposited rocks of bedrock at the highest point. This core was then overlaid by a further deposit of compacted Reddish Brown Earth mixed with flecks of bedrock. This second phase extended into pit ACRS 4A and the visible dimensions of the rampart became at least 7.34 m wide and 2.45 m high. The third phase appears to be a depositional or erosion feature rather than a construction feature. The clay and sand deposit appears to be a mixed wash from the erosion of the rampart which has collected at the inner toe. The deposit is 0.97 m at its thickest and appears to integrate with occupational sequences of the same deposit to its north. The fourth phase consisted of a layer of ashy silt, overlying the second phase mound and the third phase wash. The deposit was 1 m thick and thus increased the visible dimensions of the rampart to at least 3.30 m high and 9 m wide. The inner toe of the rampart joined occupational deposits of the same macro-context to its north. Phase five saw an additional 1.10 m height added to the rampart, making the total height at least 4.40 m. The deposit consisted of brickbats and decomposed brickbat material. The sixth phase levelled and 52

The Fonifi,cations spread the remains of the brick-built rampart and added an additional 3.5 m thick deposit of ashy silt. The enlarged rampart, now 7.9 m high, had a wall constructed at the new centre, some 10 m north of the phase one mound. The wall's foundation consisted of an ashy-silt core, containing occasional ashlar slabs, faced with two parallel lines of ashlar slabs. The foundations were 3. 6 m wide and were preserved to a height of eight slabs (1.4 m high). The presentation trench cleared a length of 24.5 m of ashlar and brick walling. The Citadel's southern gateway was identified at the westernmost point of the trench flanking the eastern side of the modem Sanghamitta Mawatha. The gatehouse, built of brickbats on an ashlar foundation, was located centrally across the rampart and was 9.06 m long. It was not possible to locate the eastern side of the gate because of the metalled nature of the road and a row of houses on its western edge. It appears to have been divided by two outer walls and two inner walls into three cells: the southern cell was 2. 73 m long, the central cell 3.6 m and the northern cell 2. 73 m long. The position of one of the gates was identified from an in situ ashlar slab with a worn socket. The wall appears to have collapsed and had been partially robbed for building materials. Two badly damaged stone bulls were found in the debris of the wall. A seventh and final construction phase added a layer of grit to the rampart, completely covering the earlier ashlar and brickbat wall. The grit layer was between 1.30 m and 0.20 m thick and gave the rampart a height of 8.1 m. A short central alignment of ashlar slabs was found near the surface, perhaps marking the final defensive wall of the Citadel. It became clear that the sequence of macro-contexts from the ACRS pits can be correlated with the sequence from many of the Citadel sondages. This correlation may be used to help us date the various phases of rampart construction from the carbon dates and artefacts already recovered from the earlier excavations. The primary phase of rampart construction was a mound built of compacted Reddish Brown Earth and bedrock fragments on its top (Fig. 48). Reddish Brown Earth is found above the basal gravels and bedrock in the Citadel as either a sterile deposit or an occupational deposit, cut by postholes and with finds of artefacts. The earliest sedentary or semi-sedentary occupation of site ASW2 was during structural period K, finds from which place it in the peninsular Indian Iron Age techno-complex, while radiocarbon results suggest a date of between the ninth and the mid-fifth centuries BC. Finds from the second structural period, J, have been dated to between circa sixth and midfourth centuries BC (Coningham et al. 1996). The primary core of the rampart yielded no artefactual remains or charcoal samples and thus is almost impossible to date accurately. Although both structural periods K and J and the rampart core are cut into layers of Reddish Brown Earth or made from unmixed

Reddish Brown Earth, this does not necessarily give a date of between circa eighth and mid-fourth centuries BC for the rampart. As stated above, the Reddish Brown Earth is also found in the Citadel sequence and elsewhere in Anuradhapura as a sterile or natural soil. It is very possible that if the rampart was constructed in a following structural period, its line would be outside the contemporary settlement and thus would involve the excavation and mounding of artefactually sterile deposits. Such sterile Reddish Brown Earth deposits were observed by the first author in the vicinity of the Rajaratta Hotel in Anuradhapura in a freshly cut, 2 m deep pond in 1992. Phase two appears to consist of a mixture of Reddish Brown Earth and the phase four silty ash. It differs from the primary core in that the presence of fragments of kiln-fired tile and the absence of kiln-fired brick attributes the deposit to ASW2 structural period I. Phase three appears to be a talus of clay and sand overlying the Reddish Brown Earth and silty-ash deposit; it may even represent a cleaning out of the moat. Phase four of the rampart consisted of a raising of the rampart height and a broadening of its base; these works were effected with an ashy-silty deposit. The sixth structural period at ASW2, I, represented a structural watershed with the replacement of round buildings with cardinally orientated square ones. The soil matrix also changes from the humus-rich Reddish Brown Earth to an ashy-silty soil identical with that used in the fourth phase of rampart construction. Finds characteristic of structural period I were also recovered from the fourth rampart deposit, including a carnelian ring, a fragment of natural glass, an amethyst bead and Rouletted ware. ACRS phases two, three and four can be interpreted as being deposits contemporary with ASW2 structural period I and can thus also be allocated a date of between circa mid-fourth century BC and the very beginning of the second century BC (Coningham et al. 1996). This date appears to corroborate the chronicle's record of the re-foundation of Anuradhapura as a royal capital by Pandukabhaya, grandfather of King Devanampiya Tissa (r. 250-210 BC). The Rajavaliyastates that 'he cleared a piece of ground, four gaw in length and the same in breadth, rooted out the trees, made streets, and constructed other works. He also built a rampart 16 gaw (in extent)' (Raj.22). In the Rajavaliya's glossary one gaw is calculated as one fourth of a yoduna, which itself represents 16 miles (Raj.vii). It thus appears that the extent described must be an exaggeration, since it is estimated that the ramparts enclose some 100 hectares. The fifth rampart phase consisted of the capping of the earlier ramparts with a brick superstructure or wall, although it appears that much of it was levelled to provide a foundation in the succeeding construction phase. The very large size of bricks appears to suggest that this phase can be correlated with the use of such brickbats at ASW2 in structural phases G and F. Structural phase G has been dated between circa the first quarter of the third century BC and the latter half of the first century AD, while structural period F can be assigned a date of between circa AD 200 and 600 (Coningham et al. 1996). The sixth rampart phase consisted of a further enlargement of the fortifications, a higher and wider rampart, and an ashlar and brickbat wall 53

Anuradhapura:The Site above (Fig. 49). The ashy-silt deposit used to raise the rampart appears to correlate to the ashy-silt occupation levels of ASW2's structural periods C, D, E and B. These four periods can be dated to between circa seventh and thirteenth centuries AD. The finds from the ACRS pits for this phase confirm the attributed date: the artefacts recovered included glass, West Asian ceramics, East Asian ceramics, glass bangles and later glass beads. The discovery of this monumental ashlar wall helps us to understand the reason for the hundreds of robber pits cut into the Citadel's earlier levels, obviously in order to recover building materials for the wall. The seventh phase, consisting of a further raising of the rampart over the collapsed wall of phase six, is most difficult to date. This late grit deposit cannot be identified in any of the Citadel excavation pits and does not include any diagnostic finds. The phase can be interpreted as either of two depositional features. Firstly, it may represent an attempt to repair the collapsed defences, perhaps carried out during one of the many attempted restorations of Anuradhapura by Polonnaruva-period rulers, Vijayabahu I (r. AD 1055-1110) (Cvs.58.59), Parakramabahu I (r. 1153-86) (Cvs.74.1-14), Parakramabahu II (r. 1236-70)(Cvs.87 .66) and Vijayabahu N (r. 1270-72) (Cvs.88.83). Secondly, it may represent the spoil thrown up by Henry Parker's irrigation ditch, which was cut along the line of the old moat in 1873.

Early Historical walled urban complex in South Asia was Dhanyakataka on the River Krishna, although the fortifications of Banavasi in Karnataka may be proved to be of Early Historic date (IndianArchaeology:A Review[IRA] 1971: 29). With the extension of the distribution to Sri Lanka, it is now obvious that the second South Asian emergence of complex societies and urbanism was not just the result of Mauryan imperial conquest, nor purely a northern phenomenon. Anuradhapura can now be added to the list of major Early Historic central places, proving that the distribution of these sites does extend outside the perimeters of North India and indeed the Mauryan empire. The early date of the city's fortifications also suggests that it was established as a major settlement before, according to the Mahavamsa,Emperor Asoka sent his son Mahinda to convert the island. As discussed in a preliminary note elsewhere (Coningham 1993), it is possible to use this new data from Anuradhapura to re-examine the possible factors behind the presence of the early fortifications at the Citadel. Defence appears to have been, logically perhaps, one of the earliest explanations for the massive Early Historic ramparts of South Asia. Certainly there can be little doubt that the walls, gateways and bastions of Sisupalgarh in Orissa were defensive. Indeed Wheeler, an experienced military man, classified the site as a fortress-town (Wheeler 1959: 134). Allchin has added further corroboration to this hypothesis for the emergence of the fortified city by stating (Allchin 1989: 4) that: . .. as the construction of these ramparts coincides with the period of emerging cities and states, and of the internecine warfare, the matsya nyaya of the Sanskrit apothegm, when state swallowed up state, until Maghadha emerged as a single overall political power, the thesis that defense was primarily against man, even though to a lesser extent against animals and floods, seems most plausible.

4.9 Conclusion All four surveys successfully achieved their aims. It became very clear, however, that the surveys were far more useful when used in combination rather than applied in isolation. The surface survey and resistivity meter identified the ashlar wall, while the proton magnetometer, resistivity meter and soil auger identified the moat. As noted above, because we wanted to test the feasibility of using geophysical prospection in Sri Lanka we selected transects and areas which should have resulted in the recording of the moat, easily identifiable as a massive positive anomaly. It is clear from selected sectors that this has largely been the case. Our second aim was to identify the course of the wall around the entire site in order to draw attention to its course for preservation and to protect it from being built on or from becoming a quarry for building materials. We have completed a plan of these results and sent this to the Archaeological Survey Department. Thirdly, we wanted to plan the course of the defences in order to estimate the area enclosed by them. Initially we had hypothesized that, as the walls represented a fairly late construction phase, they would have enclosed only part of the site, reflecting a postulated decrease in the population of the city. However it is now clear that they enclose the entire 100-hectare site. Until the discovery of the Anuradhapura rampart, dated to between ca. mid-fourth century BC and the last quarter of the third century BC, the most southerly

Mate has criticized the thesis that defence was the prime motive for the construction of these fortifications because there was no provision of a parapet and because the gentle slope of the outer face of the earliest examples made them vulnerable to attack (Mate 1970). He interprets the moats as diversion channels to ease rivers in spate and bypass the cities, rather than as part of a formally planned defensive complex. The absence of parapets on the ramparts of Ujjain, Kausambi and Rajghat is put forward as evidence to support this thesis, and he suggests that parapets were only built in the latter part of the first millennium BC. This conclusion appears to ignore the factor of archaeological survival. It is unlikely that early parapets will survive, as they will be eroded or levelled as the underlying rampart is utilized as a solid foundation for further constructional additions to the walls. Following his critique of the defensive theory, we have seen that Mate replaced it with another prime mover, that of the rampart as a flood barrier or embankment. The early archaeological levels at Hastinapura may provide some evidence for this theory. At the end of period II, characterized by finds of Painted Grey Ware, the 2.6 m high 54

The Fonifications settlement mound was partially washed away by a great river flood. This natural disaster led to the abandonment of the site and evidently 'must have entailed enormous loss oflife and property' (Lal 1955: 15). Further evidence for this theory may be found in the sequence of Rajghat's 10 m high clay rampart. The excavator, Narain, stated that 'a series of alternating deposits of sand and silt against the toe of the rampart indicated that it has been breached several times by heavy floods, which had affected some portions of the habitation' (/AR 1961: 37). Banerjee, the excavator of Ujjain, interpreted the addition of a timber framework to the early rampart of period I as a measure to protect a damaged section of wall from river erosion. However, Erdosy has noted that, as the framework was located on the inward bend of the river, it protected the city from erosion, not flooding as Mate had hypothesized (Erdosy 1988: 114). It may also be noted that there are a number of examples of early walled sites in areas not affected by flooding. Rajgir, for example, was equipped with a defensive rubble wall running along the tops of the surrounding hills (Ghosh 1951: 66). Erdosy accepts Mate's critique of the defensive prime mover but also criticizes the latter's flood-barrier hypothesis as inadequate to explain the sheer monumentality of a number of the fortifications (Erdosy 1988: 114). He also calculates that the rampart at Ujjain would have taken a labour force of 20,000 men over 250 days to complete (ibid.). In view of the vast expense incurred he appears to favour a symbolic prime mover and comments (ibid.) that: Mumford's stress on the symbolic significance of city walls, later developed by Wheatley into a contrast between sacred (urban) and profane (rural) space, provides the best explanation. Cities can thus be viewed as attempts to recreate the universe in microcosm, which needed explicitly symbolic protection in the shape of the outsized ramparts. Thus the city represents a model of the universe, the walls of the city represent the boundary of the universe and - by extension - the king represented the king of the universe! An attempt to allocate a single prime mover or function to the Citadel ramparts at Anuradhapura appears to be rather pointless as there are examples which support all of these factors. The rampart and moat doubtless functioned as part of a defensive unit. The Sinhalese chronicles document warfare from the earliest times, either against indigenous inhabitants, among the Sinhalese themselves, or against foreign expansionists and adventurers. The newly arrived Vijayan adventurers are thus recorded as having fought and defeated the native Yakkhas in order to settle the island safely (Mvs.vii.36-38). A few generations later the Sinhalese are recorded as having successional wars. Pandukabhaya, grandfather of King

Devanampiya Tissa (r. 250-210 BC), thus had to defeat eight uncles before he could claim kingship (Mvs.x.64-72). The chronicles also record that two South Indian adventurers usurped King Suratissa' s throne during the last part of the second century BC (Mvs.xxi.10-11). It may be concluded that in order to retain, or obtain, kingship, a strong army and fortress were prerequisites! The function of a rampart as a flood embarkment also appears to be satisfactorily supported by examples of natural disasters in the recent history of the island. The Dry Zone of northern and southern Sri Lanka has periodic wet-season cyclones which, in combination with heavy rains, have caused tanks to burst and rivers to flood. In December 1957 the New Town of Anuradhapura was flooded under some 2-3 m of flood water when the Malvatu Oya rose 9 m while in spate, and Parker records that in 1897 over 1 m of rain fell in just 27 hours (Parker 1909: 369). Such examples give evidence that flood barriers would have been very necessary and could have utilized the simple tank embankment technology available at that time. The early symbolic function of the rampart at Anuradhapura is more difficult to evaluate, partly because of the limited nature of the excavations at the Citadel, although there are many later conspicuous examples in the island (Coningham 1993). However we may rely, with caution, on the description of the refoundation of the settlement of Anuradhapura by King Pandukabhaya after his coronation, as documented in the Mahavamsa (Mvs.x.73-102). As discussed in Chapter 3.3 above, the king consulted a soothsayer and a site specialist before constructing the city and allocating different social groups and structures to specific loci. The very fortifications and urban plan may have been mnemonic of daily life or rather - as Thapar suggests (1984: 91) - symbolic in that: The fortifications enclosed the urban settlement and separated it from the surrounding areas . . . thus demarcating the urban from the rural ... Fortifications also served to segregate excluded social groups such as the Candalas who lived in villages in the vicinity. Part of the function of the early ramparts was undoubtedly to exclude enemies, prevent flooding and act as a symbol of the king's ritual and cosmic role. However, one factor appears to have been omitted from this list of multivariants - the protection of crops. One of the most obvious features of early historic cities is the enormous hectarage enclosed. Erdosy calculates that 11 Gangetic examples covered over 100 hectares (Erdosy 1988: 134). It is highly improbable that in the early phases of these settlements all of the enclosed area was occupied by housing. Erdosy calculates that Kausambi's defences were erected c. 500 BC and enclosed 250 hectares. However, according to his surface survey, only 50 hectares were occupied in period II (600-350 BC) (ibid.: 60) and only 150 hectares in phase III (350-100 BC) (ibid.: 72). Erdosy's calculation thus leaves a huge percentage of land within the ramparts unoccupied by settlement: 80 percent in period II and 40 percent in period III. It is highly probable that much of this land was occupied by market or kitchen gardens and groves of fruit trees. This pattern also appears to have been detected at 55

Anuradhapura:The Site Anuradhapura. The ramparts of structural phase I there enclosed an area of 100 hectares, two thirds of which were occupied, leaving a third unoccupied. Contemporary subsistence strategies in North Central Province generally rely upon three main traditional techniques: tank-irrigated rice, chena(slash and bum), and the cultivation of garden plots (Leach 1961). All these options necessitate some degree of protection for the crops, especially in the more isolated settlements in jungle areas. When crops of irrigated rice-fields are near harvesting, a watch is normally kept day and night to ensure that they are not destroyed either by wild pigs and deer or by domestic buffalo and cows. Chenaor newly cleared areas are far more difficult to protect, but they are vital because they supplement the mainly rice-based diet. Baker recorded that Korrakan, maize, Indian com, millet and pumpkins were grown on such land (Baker 1855: 35). Although most areas of chena are fenced, pigs can dig under them, deer can jump over them and elephants can trample them down. Often, in more remote areas, farmers still build platforms in tall trees where they light fires and shout and shake rattles all night to protect themselves and their crops from wild animals. Village gardens also supplement the rice diet in the form of the yields from coconut palms and fruit trees, as noted by Knox in the seventeenth century (Knox 1911: 141). The former, if unprotected, were often knocked down by elephants trying to reach the succulent tops (Baker 1855: 46). Indeed, elephants were such a menace to the economy of the island in the nineteenth century that the Government offered 10 shillings an elephant tail in certain areas, although this was soon abolished because the Government quickly found the bounty too expensive (ibid.: 67). It is worth noting that it is well within the capacity of an adult elephant to eat over 1000 pounds of fodder in an hour (Deraniyagala 1955). Domestic livestock was also at risk from jackals and leopards, and the English inhabitants of Nuwara Eliya suffered badly from the latter (Baker 1955: 59). In view of the above evidence it appears no surprise that when Robert Knox acquired some land in the hill country and built a house there in 1666, the first action he took was to 'intrench it round with a ditch, and planted a hedge' (Knox 1911: 141). Even so, he records that the enclosed land was often broken

into by sambhur, wild pigs and leopards (ibid.: 26-27). His second house and land were similarly defended against wild animals, and the entrances were protected by thorn fences (ibid.: 149). Similar defensive enclosures are described in connection with a twelfth-century AD village in the Culavamsa (Cvs.66.87). We may therefore surmise that the first fortified settlement at the Citadel of Anuradhapura enclosed an area of some 100 hectares, of which slightly more than two thirds were occupied. Part of the impetus for this enormous work may have been as a symbolic barrier between order and disorder, or a defensive fortification against flooding or against inhabitants from hostile polities or settlements. However, it also provided a physical barrier which wild pigs, deer, leopards, jackals and wild elephants could not surmount, trample down or dig under to reach the kitchen gardens, fruit trees or crops that may have been planted on the remaining unoccupied third of the settlement. It is also clear that the earliest rampart at Anuradhapura represents a large investment of communal action. The ramparts run for some 2980 m, are some 2.10 m high and have an estimated width of 8 m (Fig. 50). Their volume can be calculated at around 50,064 m3• We can calculate the number of man-days taken to build the rampart by using a rate of 0. 58 m3 per man-day, based upon observations of nineteenth-century canal digging (Erdosy 1988: 113). They are equal to 86,317.241 mandays or, if one assumes that the rampart was built well within the dry season when excess labour was available, one may calculate that it would have taken a postulated workforce of 575 a total of 150 days. The mobilization of large numbers of people is also suggested from the construction of large tanks for irrigated rice and the watering of growing numbers of livestock and people. The change to the environment and the drop in the water-table is archaeologically visible: the shallow watering holes of structural periods J and K were replaced during period I by deep wells cut through the underlying deposits and into the bedrock. These collective works mark Anuradhapura as the primate city and illustrate the island's earliest example of the ability to mobilize a large labour force in the field. They thus mark the beginning of complex societies in Sri Lanka which culminated in the classical Anuradhapura period.

56

~,I

,.·-·-·r---·• .• . :

~---

ha;gamttI,if's.--

H

./ ..·······.r • ....-•······ . · ..:: ~!J,.•.·,r ft.. ···: '

-~ J:IAY.A1

:'I!!

- ~-, "'l ••TI~ .. ..,!!\. ti""'"'·::.::= ~\\- ~r.~:-r ..,., . \1 . .__ \ llt'f:lf:: ... .:.. .... ,

Mlf;,~ ./ fl

/

A}1,,

l

(.

li:iJ

I

I

'1

i,___ t,·-;;~'l[\'tlJldu Temple~JI h 11 n //· -,t::=--=-··;:::::;;:;;"

/I \\, . / --............... >I

1

1

....

.

I

~

I :: I,

V,

-..J

i

I

/1

. )\;) ==:-=-:~~~;.~:~. ". ~=~==~::.~::.-==;~=::::::111.. ,r tPalhamaCpi,v,r,) _.

•

1~

I ',

1~1

. : fL• • I i :l .._·,;;~;;,;;;,;~--

',,,

~

\

\,,: ' rr ::i 1 •t(

.....\

,Ji

',,,,

,

l i iJ 1 &.......... .

Ji , :

\•.

/ '\ ~I

...........

~.BrU;y,,

I

1' ....,.

,

ll

..,..._. ~-B»tdfle

:=::::::::.::.-::~,

"'

~ 'Si

'

~

-l'al

g

...-.:'

~·

/~.,,/

- tf

BA 5 A VA I< K UL A M

It-------11

(.

A8HAYA

~

Vl'f Pl

)

. "'

-,.,.·

I•

,......,-,. ..,II • lo' i

d•

// ',

11--.:

,..~ ..l

' '

~aM

400 metres

j:: •.;

i,

I,~

•

1

i.";

. ;! ,0• l • t. ·-.

~.

ANU.R1ADHAPUR r;-.-.1

~ ..... (I:>

.....

"' "'

1557 nw 1558 nw 1559 nw 1560 nw 1561 nw 1562 nw 1563 nw 1564 nw 1565 nw 1566 nw 1567 nw 1568 nw 1569 nw 1570 nw 1571 SW 1572 SW 1573 SW 1574 SW 1575 SW 1576 SW 1577 SW 1578 SW 1579 SW 1580 SW 1581 SW 1582 SW 1583 SW 1584 nw 1585 SW 1586 SW 1587 SW 1588 SW 1589 SW 1590 SW 1591 SW 1592 SW 1593 SW 1594 SW 1595 SW

2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91

10yr 3/4

70c.30s

10yr 3/3

40st.40c.20s

10yr 3/3

40st.40c.20s

10yr 3/3

40st.40c.20s

10yr 3/6

40st.40c.20s

10yr 3/3

40st.40c.20s

10yr 3/4

40st.40c.20s

10yr 3/3

70s.30st

10yr 3/6

60c.40s

10yr 3/4

70c.30s

10yr 3/4

14x11cm

22.2cm

10cm

14.5cm

10cm

14cm

15cm

8cm

20cm

20.5cm

15x10cm

4cm

15cm

6cm

27x15cm

35.5cm

28x15cm

11cm

13cm

6.5cm

13cm

11cm

10cm

7cm

8cm

10cm

13cm

12cm

11cm

8cm

7cm

5cm

20cm

5.5cm

11cm

19.5cm

9cm

7cm

9cm

11cm

65c.35s

10yr 3/3

60c.40s

10yr 3/4

70c.30s

10yr 3/4

65c.35s

10yr 3/4

70c.30s

10yr 3/3

55c.45s

10yr 6/8

100s

10yr 3/4

70c.30s

10yr 3/4

70c.30s

b1407 c1496 fill of 1557 b1407 c1496 fill of 1559 b1407 c1496 fillof1561 b1407 c1496 fill of 1563 b1407 c1496 fill of 1565 b1407 c1496 fill of 1567 b1407 c1496 fill of 1569 b1407 c1496 fill of 1571 b1407 c1496 fill of 1573 b1407 c1496 fill of 1575 b1407 c1496 fill of 1577 b1407 c1496 fill of 1579 b1407 c1496 fill of 1581 b1407 c1496 fill of 1583 b1407 c1496 fill of 1585 b1407 c1496 fill of 1587 b1407 c1496 fill of 1589 b1407 c1496 fill of 1591 b1407 c1496 fill of 1593 b1407 c1496

posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole

XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII

J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1

~ 'IS ~

~ txi

N 0 0

1596 SW 1597 se 1598 se 1599 se 1600 se 1601 se 1602 se 1603 se 1604 se 1605 se 1606 se 1607 SW 1608 SW 1609 se 1610 se 1611 se 1612 se 1613 se 1614 se 1615 seine 1616 all 1617 seine 1618 ne 1619 ne 1620 ne 1621 ne 1622 ne 1623 ne 1624 ne 1625 ne 1626 ne 1627 ne 1628 ne 1629 ne 1630 ne 1631 ne 1632 ne 1633 ne 1634 ne

2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 2.9.91 5.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91

10yr 314

65c.35s

10yr 313

70c.30s

10yr 314

70c.30s

10yr 414

60c.40s

10yr 518

80s.20c

10yr 314

65c.35s

10yr 416

55s.45c

10yr 314

65c.35s

10yr 314

65c.35s

10yr 512 10yr 212 7.5yr 314 10yr 212

70s.30c 100c 70st.30s 100c

10yr 314

28x22cm

11cm

12x10cm

4.5cm

12cm

11cm

20cm

14cm

15cm

7cm

19x10cm

11cm

10cm

17cm

22cm

34cm

13x11cm

4.5cm

13cm

5cm

8cm

5cm

9cm

5cm

16cm

31cm

9cm

5cm

12x9cm

12.5cm

19x15cm

20cm

12cm

4cm

18x9cm

13cm

65st.35s

7.5yr 312

60st.40s

10yr 414

85c.15s

10yr 514

100c

10yr 314

85c.15s

7.5yr 314

70st.30s

7.5yr 518

80s.20c

5yr 313

75st.25s

fill of 1595 b1407 c1496 fill of 1597 b1407 c1496 fill of 1599 b1407 c1496 fill of 1601 b1407 c1496 fill of 1603 b1407 c1496 fill of 1605 b1407 c1496 fill of 1607 b1407 c1496 fill of 1609 b1407 c1496 fill of 1611 b1407 c1496 fill of 1613 b1496 a1616 b1496 a1714 '=1615 b1615 c1616 fill of 1618 b1615 c1616 fill of 1620 b1615 c1616 fill of 1622 b1615 c1616 fill of 1624 b1615 c1616 fill of 1626 b1615 c1616 fill of 1628 b1615 c1616 fill of 1630 b1615 c1616 fill of 1632 b1615 c1616

posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill redeposited? old land surf redeposited? posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole

XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII XIII X

J1 J1 J1 J1 J1 J1 J1 J1 J1 J1 J1

J1 J1 J1 J1 J1 J1 J1 J1

IX IX IX IX IX

§= ~

:::::

~

vm X IX IX IX IX IX IX IX IX IX IX IX IX

::i... ;::s ::::: ~

K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3

~

"' ~ ~

N

0

......

1635 ne 1636 ne 1637 ne 1638 ne 1639 ne 1640 ne 1641 ne 1642 ne 1643 ne 1644 ne 1645 ne 1646 ne 1647 ne 1648 ne 1649 nw 1650 nw 1651 nw 1652 nw 1653 nw 1654 nw 1655 nw 1656 nw 1657 nw 1658 se 1659 se 1660 nw 1661 nw 1662 se 1663 se 1664 nw 1665 nw 1666 nw 1667 nw 1668 nw 1669 nw 1670 se 1671 se 1672 nw 1673 nw

4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 4.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91

7.5yr 5/8 7.5yr 3/4

80s.20c

60s1.40s

5yr 4/6

65s.35st

6cm

3cm

14cm

11cm

17cm

12cm

14x8cm

3cm

12cm

4cm

7cm

3cm

10cm

46cm

10cm

6.5cm

13cm

12.5cm

14cm

25cm

13cm

23.5cm

10cm

4.5cm

16x12cm

24.5cm

11cm

8cm

65st.35s

7.5yr 3/4

65st.35s

7.5yr 3/2

75st.25s

7.5yr 3/4

80s1.20s

10yr 3/4

80s.20c

7.5yr 3/4

80c.20s

10yr 6/6

13cm

60st.40s

5yr 3/3

7.5yr 3/2

16cm

100s

10yr 3/6

90s.10c

10yr 3/4

75c.25s

7.5yr4/4

60c.40s

7.5yr 4/2

70c.30s

10yr 3/4

80c.20s

7.5yr 4/2

80c.20s

10yr 3/4

60c/40s

10yr 4/6

60c.40s

7.5yr 3/4

80s.20c

19cm

12cm

9cm

5cm

13cm

7cm

13cm

18.5cm

fill of 1634 b1615 c1616 fill of 1636 b1615 c1616 fill of 1638 b1615 c1616 fill of 1640 b1615 c1616 fill of 1642 b1615 c1616 fill of 1644 b1615 c1616 fill of 1646 b1615 c1616 fill of 1648 b1615 c1616 fill of 1650 b1615 c1616 fill of 1652 b1615 c1616 fill of 1654 b1615 c1616 fill of 1656 b1615 c1616 fill of 1658 b1615 c1616 fill of 1660 b1496 c1615 fill of 1662 b1615 c1616 fill of 1664 b1615 c1616 fill of 1666 b1615 c1616 fill of 1668 b1496 c1615 fill of 1670 b1615 c1616 fill of 1672

posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill old posthole? posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill

IX IX IX IX IX IX

IX IX IX IX IX IX IX IX IX IX IX IX IX IX IX IX IX IX IX

IX IX XI XI IX

IX IX IX IX

IX XI XI

IX IX

K3 K3

K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3

K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3

K3 K3 K3 K3? K3? K3 K3 K3 K3 K3

K3 K3? K3? K3 K3

~ 'IS

"' ~

~

t:,;;

N 0 N

1674 nw 1675 nw 1676 nw 1677 nw 1678 nw 1679 nw 1680 nw 1681 nw 1682 nw 1683 nw 1684 nw 1685 nw 1686 1687 1688 ne 1689 ne 1690 ne 1691 ne 1692 ne 1693 ne 1694 ne 1695 ne 1696 ne 1697 ne 1698 ne 1699 ne 1700 ne 1701 ne 1702 ne 1703 ne 1704 ne 1705 ne 1706 ne 1707 ne 1708 nw 1709 nw 1710 nw 1711 nw 1712 nw

5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91

5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91 5.9.91

7cm 10yr 3/4

75c.25s

10yr 3/4

75c.25s

7.5yr 3/4

70c.30s

10yr 8/8

100s

7.5yr 3/4

75st.25s

7.5yr 3/4

70st.30s

5yr 3/2

55st.45c

7.5yr 4/4 7.5yr 3/2 7.5yr 4/4

3cm

19x16cm

3cm

6cm

4cm

15cm

23cm

13x10cm

11cm

11cm

9.5cm

35x15cm

8cm

23x13cm

7.5cm

60st.40s 10cm

10cm

10cm

4.5cm

8x5cm

5cm

60c.40st 50c.50s

10yr 3/4

70c.30s

7.5yr 3/2

70st.30s

7.5yr 3/2

90c.10s

7.5yr 3/4

80s.20c

7.5yr 3/4

45st.55c

7.5yr 4/4

90c.10s

7.5yr 3/4

70c.30s

2.5y 6/4

100s

10cm

6.5cm

25x18cm

23cm

25cm

10cm

8x5cm

6.5cm

11cm

7cm

8cm

3cm

12cm

20cm

12cm

10cm

b1615 c1616 fill of 1674 b1615 c1616 fillof1676 b1615 c1616 fill of 1678 b1615 c1616 fill of 1681 b1615 c1616 fill of 1682 b1615 c1616 fill of 1684

posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill

IX

b1615 c1616 fill of 1688 b1615 c1616 fill of 1690 b1615 c1616 fill of 1692 b1615 c1616 fill of 1694 b1615 c1616 fill of 1696 b1615 c1616 fill of 1698 b1615 c1616 fill of 1700 b1615 c1616 fill of 1702 b1615 c1616 fill of 1704 b1615 c1616 fill of 1706 b1615 c1616 fill of 1708 b1615 c1616 fill of 1710 b1615 c1616

posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill old posthole? posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole

IX IX IX

IX IX IX IX IX IX IX IX IX IX IX

IX IX IX IX IX IX

IX IX IX IX IX IX IX IX IX IX

IX IX IX IX IX IX

K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3

K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3 K3

~

;::: ~

~

~ ~ ~

~ ~ (I) ~ ~

N

0

w

1713 nw 1714 all 1715 nw 1716 nw 1717 nw 1718 nw 1719 nw 1720 nw 1721 nw 1722 nw 1723 nw 1724 nw 1725 nw 1726 nw 1727 nw 1728 nw 1729 nw 1730 nw 1731 nw 1732 nw 1733 nw 1734 nw 1735 nw 1736 ne 1737 ne 1738 ne 1739 ne 1740 ne 1741 SW 1742 SW 1743 SW 1744 SW 1745 SW 1746 SW 1747 SW 1748 SW 1749 SW 1750 SW 1751 SW

6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91

2.5y 5/6 7.5yr 4/4

100s 60c.40s

10y 3/4

70c.30s

10yr 6/6

100s

10yr 4/6

50s.50g

10yr 4/4

70c.30s

10yr 3/4

65c.35s

10yr 3/4

60c.40s

10yr 3/4 7.5yr 3/4 10yr 3/4 10yr 3/6 10yr 3/6 10yr 3/6

15cm

6.5cm

16cm

18.5cm

16cm

7.5cm

15x10cm

3.5cm

19cm

12cm

15cm

8cm

12cm

6cm

11cm

6cm

8cm

5cm

70c.30s 60c.40s/g 65c.35s 6cm

5cm

15cm

6.5cm

9cm

3cm

10cm

5cm

5cm

3cm

7cm

2cm

11cm

4cm

9cm

3cm

14cm

6cm

17cm

7cm

65c.35s 50c.50s 75st.25c

10yr 3/4

60c.40s

7.5yr 3/2

70c.30s

7.5yr 3/2

70c.30s

5yr 3/4

70c.30s

5yr 3/4

80c.20s

7.5yr 3/2

65c.35s

fillof1712 b1616 a1811 b1616 c1714 fillof1715 b1616 c1714 fill of 1717 b1616 c1714 fillof1719 b1616 c1714 fill of 1721 b1616 c1714 fill of 1723 b1616 c1714 fill of 1725 b1616 c1714 fill of 1727 b1616 c1714 fill of 1729 b1616 c1714 fill of 1731 b1616 c1714 fill of 1733 b1616 c1714 fill of 1735 b1616 c1714 fill of 1737 b1616 c1714 fill of 1739 b1616 c1714 fill of 1741 b1616 c1714 fill of 1743 b1616 c1714 fill of 1745 b1616 c1714 fill of 1747 b1616 c1714 fill of 1749 b1616 c1714

posthole fill old land surf posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole

IX

VI VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII

VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII

K3 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2

K2 K2

K2 K2 K2 K2 K2

K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2

K2 K2 K2 K2 K2

~ '1;:j II:>

;::

~

~0:1

N 0

~

1752 SW 1753 SW 1754 SW 1755 SW 1756 SW 1757 se 1758 se 1759 nw 1760 nw 1761 SW 1762 SW 1763 nw 1764 nw 1765 nw 1766 nw 1767 SW 1768 SW 1769 SW 1770 SW 1771 nw 1772 nw 1773 se 1774 se 1775 nw 1776 nw 1777 SW 1778 SW 1779 SW 1780sw 1781 SW 1782 SW 1783 nw 1784 nw 1785 SW 1786 SW 1787 nw 1788 nw 1789 ne 1790 ne

6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 6.9.91 7.9.91 7.9.91 7.9.91 7.9.91

7.5yr 3/2

75c.25st

7.5yr 3/2

70sUc.30s

7.5yr 3/2

60c.40s

10yr 5/8

50s.50c

7.5yr 3/4

75c.25s

7.5yr 3/2

70c.30s

7.5yr 3/2

80c.20s

7.5yr 3/4

70c.30s

7.5yr 3/4

80c.20s

10yr 3/4

70c.30s

10yr 3/4

70c.30s

17cm

5cm

175x65cm

25cm

23x14cm

6cm

18cm

12cm

10cm

5cm

23x14cm

9cm

12cm

10cm

10cm

6cm

8cm

7.5yr 3/4

4cm

7cm

3cm

15x12cm

15.5cm

11cm

10cm

20cm

13cm

22cm

3cm

6cm

3cm

8cm

6.5cm

8cm

15cm

15x14cm

13cm

11cm

8cm

80c.20s

10yr 3/4

70c.30s

10yr 3/4

70c.30s

2.5yr 3/4

70c.30s

10yr 3/4

70c.30s

10yr 3/4

70c.30s

10yr 4/3

70c.30s

10yr 3/3

60c.40s

7.5yr 3/4

60c.40s

fill of 1751 b1616 c1714 fill of 1753 b1616 c1714 fill of 1755 b1616 c1714 fill of 1757 b1616 c1714 fill of 1759 b1616 c1714 fill of 1761 b1616 c1714 fill of 1763 b1616 c1714 fill of 1765 b1616 c1714 fill of 1767 b1616 c1714 fill of 1769 b1616 c1714 fill of 1771 b1616 c1714 fill of 1773 b1616 c1714 fill of 1775 b1616 c1714 fill of 1777 b1616 c1714 fill of 1779 b1616 c1714 fill of 1781 b1616 c1714 fill of 1783 b1616 c1714 fill of 1785 b1616 c1714 fill of 1787 b1616 c1714 fill of 1789

posthole fill posthole posthole fill pit pit fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill

VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII

VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII

K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2

K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2

K2

::i,.. ;:s ~

~

~

{5 ~

~ ~

"'

~ ~

N

0 Ul

1791 nw 1792 nw 1793 nw 1794 nw 1795 nw 1796 nw 1797 nw 1798 nw 1799 nw 1800 nw 1801 nw 1802 nw 1803 nw 1804 nw 1805 SW 1806 SW 1807 ne 1808 ne 1809 ne 1810 ne 1811 all 1812 ne 1813 ne 1814 ne 1815 ne 1816 ne 1817 ne 1818 ne 1819 ne 1820 ne 1821 ne 1822 ne 1823 ne 1824 ne 1825 ne 1826 ne 1827 ne 1828 nw 1829 nw

7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 7.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91

7.5yr 3/2

50c.50s

10yr 3/4

60c.40s

10yr 4/4 10yr 4/4 10yr 2/2

16x9cm

4cm

12cm

6.5cm

14cm

5cm

10cm

5cm

10cm

10cm

3cm

5cm

11cm

5cm

8cm

2cm

7cm

8cm

7cm

7cm

60c.40s 60c.40s 60c.40s

10yr 4/4

60c.40s

5yr 3/3

70c.30s

7.5yr 3/4

70c.30s

10yr4/4

60c.40s

10yr4/4 7.5yr 3/4

50s.50c 80c.20s

10yr 3/6

60c.40s

10yr 3/4

70c.30s

10yr 3/4

70c.30s

10yr 3/6

70s.30c

7.5yr 3/4

50c.50s

10yr 3/4

70c.30s

10yr 3/4

70c.30s

10yr 4/4

70c.30s

10yr 4/4

60c.40s

20cm

9cm

13cm

6.5cm

10cm

7cm

14cm

12.5cm

10cm

5cm

10cm

5cm

9cm

6.5cm

10cm

7.5cm

11cm

3cm

b1616 c1714 fill of 1791 b1616 c1714 fill of 1793 b1616 c1714 fill of 1795 b1616 c1714 fill of 1797 b1616 c1714 fill of 1799 b1616 c1714 fill of 1801 b1616 c1714 fill of 1803 b1616 c1714 fill of 1805 b1616 c1714 fill of 1807 b1616 c1714 fill of 1809 b1714 a1887 b1714 c1811 fill of 1812 b1714c1811 fill of 1814 b1714 c1811 fillof1816 b1714 c1811 fill of 1818 b1714 c1811 fill of 1820 b1714 c1811 fill of 1822 b1714 c1811 fill of 1824 b1714c1811 fill of 1826 b1714 c1811 fill of 1828

posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill old land surf posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill

VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII VII

VII VII VII VII IV V V V V V V V V V V V V V V V V

V V

K2 K2 K2 K2

K2 K2 K2

K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2 K2

~ ~

K2

~

K1 K1 K1 K1 K1 K1 K1 K1 K1 K1 K1 K1 K1 K1 K1 K1 K1 K1

b:i

~

~-

N 0

°'

1830 nw 1831 nw 1832 nw 1833 nw 1834 nw 1835 nw 1836 nw 1837 nw 1838 nw 1839 nw 1840 nw 1841 nw 1842 nw 1843 nw 1844 nw 1845 nw 1846 nw 1847 nw 1848 nw 1849 nw 1850 nw 1851 nw 1852 nw 1853 nw 1854 se 1855 SW 1856 SW 1857 SW 1858 SW 1859 SW 1860 SW 1861 ne 1862 ne 1863 ne 1864 ne 1865 nw 1866 nw 1867 nw 1868 nw

9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91

10yr 6/6

100s

10yr 3/4

60c.40s

10yr 4/4

60c.40s

10yr4/4

60c.40s

10yr 3/6

65c.35s

10yr 3/6

80c.20s

10yr 3/6

70c.30s

5yr 4/4

4x8cm

3cm

7cm

6cm

55x17cm

8cm

10cm

6cm

9cm

3cm

9cm

8.5cm

5cm

5.5cm

11cm

4.5cm

10cm

6cm

80c.20s

10yr 3/6

70c.30s

10yr 2/2

70c.30s

15cm

10yr 3/4

8cm

23cm

6.5cm

17cm

8cm

15cm

10cm

10cm

4cm

16cm

3cm

80c.20s

7.5yr 3/4 10yr 4/4

70c.30s 75c.25s

7.5yr 3/4

70c.30s

5yr 3/4

55c.45s

7.5yr 3/4

60c.40s

10yr 4/4

90s.10s

7.5yr 3/4

50s.50c

7.5yr 3/4

75c.25s

7.5yr 3/2

80c.20s

16x13cm

8cm

13cm

6.5cm

12cm

6.5cm

17cm

6.5cm

b1714 c1811 fill of 1830 b1714c1811 fill of 1832 b1714 c1811 fill of 1834 b1714 c1811 fill of 1836 b1714c1811 fill of 1838 b1714 c1811 fill of 1840 b1714 c1811 fillof1842 b1714 c1811 fill of 1844 b1714 c1811 fill of 1846 b1714 c1811 fill of 1848 b1714 c1811 fill of 1850 b1714 c1811 fill of 1852 fill of b1714 c1811 fill of 1855 b1714c1811 fill of 1857 b1714 c1811 fill of 1859 b1714c1811 fill of 1861 b1714 c1811 fill of 1863 b1714c1811 fill of 1865 b1714c1811 fill of 1867

posthole posthole fill posthole posthole fill slot? slot? fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill pit fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill posthole posthole fill

V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V V

K1 K1

K1 K1 K1 K1 K1 K1 K1 K1 K1 K1 K1

K1 K1 K1 K1 K1 K1 K1 K1 K1 K1 K1

K1 K1 K1 K1 K1 K1 K1 K1 K1

K1 K1 K1 K1 K1

::i.. ;::s l:::

(j

~

'G

;::

~

~

(1:,

~

N

0 ....J

1869 se 1870 se 1871 ne 1872 ne 1873 nw 1874 nw 1875 nw 1876 nw 1877 nw 1878 nw 1879 nw 1880 nw 1881 SW 1882 SW 1883 SW 1884 SW 1885 se 1886 se 1887 all 1888 all

9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 9.9.91 10.9.91 10.9.91 10.9.91 10.9.91 10.9.91 10.9.91 10.9.91 10.9.91 10.9.91 10.9.91 10.9.91 10.9.91 10.9.91 10.9.91

7.5yr 3/2

70c.30s

10yr 4/3

60c.40s

10yr 4/4 2.5yr 4/4

7.5yr 2.5/4

70c.30s

7.5yr 3/4

15x8cm

4.5cm

12x7cm

4cm

30cm

10cm

18cm

4.5cm

20cm

3.5cm

65x60cm

25cm

40cm

9cm

45x40cm

36.5cm

70c.30s 50c.50s

7.5yr 3/4

4.5cm

80c.20s

7.5yr 4/4

7.5yr 3/2

17x11cm

80c.20s 80c.20s 75c.25s 100g 100rock

b1714 c1811 fill of 1869 b1714 c1811 fill of 1871 b1714 c1811 fill of 1873 b1811 c1887 fill of 1875 b 1811 c1887 fill of 1877 b1811 c1887 fill of 1879 b1811 c1887 fill of 1881 b1811 c1887 fill of 1883 b1811 c1887 fill of 1885 b1811 a1888 b1888

posthole posthole fill posthole posthole fill posthole posthole fill pit? pit? fill posthole? posthole fill posthole? posthole fill pit? pit? fill pit? pit? fill pit? pit? fill basal gravel bedrock

V V V V V V III III

K1 K1

III III

K K K K K

III III III III III III III

III II I

K1 K1 K1 K1 K K

K K K K K

t

'IS

~

~ O;:j

APPENDIXC RADIOCARBON AGES OF MEASURED SAMPLES FROM ASW2

N

0 \0

Phase

Context

Sample reference

Radiocarbon age

Sample material

Archaeological context

G5 G3 G2

340 632 615

BM-2781 Beta-48939 Beta-48938

1950+/-60 BP 1950+/-60 BP 2130+/-60 BP

Charcoal Charcoal Charcoal

Fill of posthole 341 Foundation construction/levelling Bulked sample from levelling/occupation floor

H H H H

735 692 718 721

BM-2878 Beta-48937 Beta-48936 Beta-48935

1960+/-80 2250+/-60 2280+/- 70 2230+/-90

BP BP BP BP

Charcoal Charcoal Charcoal Charcoal

Basal Basal Basal Basal

18 18 17 15 14 14

728 812 834 901 905 914 1097 1112 1113 1173 1173

Beta-48934 Beta-48933 Beta-48932 Beta-48931 Beta-48930 BM-2876 Beta-48928 Beta-48927 Beta-48926 Beta-48925 BM-2877

2000+/-80 2220+/-80 2160+/-60 2320+/-60 2390+/-60 2200+/-45 2290+/-90 2110+/-60 2150+/-60 2150+/-50 2310+/- 70

BP BP BP BP BP BP BP BP BP BP BP

Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal

Basal fill of post slot 737 Basal fill of post slot 811 Bulked sample from levelling/occupation floor Fill of posthole 900 Carbonised timbers under roof collapse Carbonised timbers under roof collapse Basal fill of oven/furnace 1096 Basal fill of stoke hole 1111 Basal fill ofoven/furnace 1109 Basal fill of oven/furnace 1148 Basal fill of oven/furnace 1148

12 JI

1236 1291 1291 1175 1382 1342 1417 1496

Beta-48919 Beta-57702 Beta-48918 Beta-57701 Beta-48923 Beta-48924 Beta-48922 Beta-48921

2160+/-60 BP 2300+/-90 BP 2240+/-80 BP 2380+/- 70 BP 2140+/-70 BP 2320+/- 70 BP 2230+/-60 BP 2310+/-100 BP

Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal Charcoal

Basal fill of oven/furnace 1235 Basal fill of oven/furnace 1235 Basal fill of oven/furnace 1235 Bulked sample from levelling/occupation floor Fill of burial? pit 1371 Basal fill ofoven/furnace 1341 Fill ofposthole 1416 Bulked sample from levelling/occupation floor

K3 K2 Kl

1616 1714 1811

Beta-48920 Beta-48917 Beta-48916

2490+/-60 BP 2360+/- 70 BP 2550+/-80 BP

Charcoal Charcoal Charcoal

Bulked sample from levelling/occupation floor Bulked sample from levelling/occupation floor Bulked sample from levelling/occupation floor

13 12 12 II II

14 14 14 14 J3 J3

i i.

fill fill fill fill

of of of of

oven/furnace oven/furnace oven/furnace oven/furnace

735 733 735 738